,
'
Electricity from Photovoltaic Solar Cells
Flat-Plate Solar Array Project
10 Years of Progress
October 1 985
••
,United States
Department of Energy
.JPL
Jet Propulsion Laboratory
California Institute of Technology
NI\SI\
National Aeronautics and
Space Administration
Photovoltaic Module Progress
75
MODULE PRICE
~
«
LU
;01<
a...
10°.
>u
ff-
«
Z
~
LU
u
<f}
5tt
LU
U
(f)
a::
«
LU
>-
LU
a:
a...
5
1975
1985
Flat or nonconcentrating module prices have dropped
as module efficiencies have increased. Prices are in
1985 dollars for large quantities of commercial products.
1975
1985
Typical module lifetimes were less than 1 year but
are now estimated to be greater than 10 years.
(Ten-year warranties are now available.)
Technology advancement in crystalline silicon solar cells
and modules (non-concentrating).
Union Carbide Corporation (UCC) funded the now
operational silicon refinement production plant with
1200 MTlyear capacity. DOEIFSA-sponsored efforts
were prominent in the UCC process research
and development.
The automated machine interconnects solar cells
and places them for module assembly. The secondgeneration machine made by Kulicke and Soffa was
cost shared by Westinghouse Corporation
and DOEIFSA.
Module evolution is represented by typical modules from
the series purchased by DOEIFSA between 1975 and 1985.
The modules were designed and manufactured by industry
to FSA specifications and were evaluated by FSA.
FRONT COVER PHOTO:
More technology advancements are shown on
inside of back cover. Use of modules in photo voltaic
power systems are shown on outside of back cover.
The Block I module (1975) being
held in front of four Block V modules
represents the progress of
a la-year cooperative
industryluniversitylDOEIFSA
effort.
Electricity from Photovoltaic Solar Cells
Flat-Plate Solar Array Project
of the
U.S. Department of Energy's
National Photovoltaics Program
10 Years of Progress
October 1985
Elmer Christensen
. United States
Department of Energy
i
"
.lPL
Jet Propulsion Laboratory
California Institute of Technology
NI\S/\
National Aeronautics and
Space Administration
The Role of Photovoltaics:
An Historical Perspective
The world's demand for more energy continues,
and will continue in the foreseeable future. The
economic growth of developing nations depends upon
increasing energy supplies. In technically advanced
countries, long-range economic growth depends on
the availability of inexpensive energy. Historically, in
the United States, the main source of energy has
changed from wood to coal to petroleum, and it is
inevitable that change will continue as fossil fuels are
depleted. Within a lifetime, it is expected that most
U.S. energy will come from diverse sources, including
renewables, instead of from a single type of fuel.
There are many indications that the percentage of the
total consumed energy supplied by electrical power
will continue to increase.
objective of reducing the cost of manufacturing solar
cells and modules. This effort, which was assigned to
the Jet Propulsion Laboratory (JPL) in Pasadena,
California, evolved from more than a decade and a
half of spacecraft PV power-system experience at JPL
and from recommendations of a conference held in
1973 at Cherry Hill, New Jersey, on Solar
Photovoltaic Energy.
After about 2 years of technical progress and
economic analyses by the Flat-Plate Solar Array
(FSA) Project, a growing community of people
became convinced that the Project's goals were
realistic and that PV power systems had the potential
of being competitive with those based on fossil fuels.
Additional federally supported efforts, sponsored by
the U.S. Department of Energy (DOE), covering all
aspects of PV power-system R&D, were initiated at
the Solar Energy Research Institute (SERI), Sandia
National Laboratories, and other organizations. These
broadened efforts included concentrator modules, PV
systems studies, application demonstrations, and
research on cell materials other than silicon and
devices such as thin-film cells. Most of the DOE funds
have sponsored R&D in private organizations and
universities. The result has been an effective
Government-university-industry team that has
cooperated to advance PV technology rapidly.
The use of photovoltaics (PV) is a promising way to
generate electricity that could provide significant
amounts of power in the years to come. It is a form of
solar energy that uses solar cells to generate
electricity cleanly and reliably, directly from sunlight
without moving parts. PV power systems are simple,
flexible, modular, and adaptable to many different
applications in an almost infinite number of sizes and
in many environments. Although photovoltaics is a
proven technology that is cost-effective for hundreds
of small applications, it is not yet cost-effective for
large-scale utility use in the United States. For
economical widespread use, the cost of generating
PV power must continue to be decreased by reducing
the initial system cost, by reducing land or area
requirements, and by increasing the operational
lifetime of its systems.
In parallel, a small but fast-growing terrestrial PV
industry evolved, with products that were
economically competitive for small stand-alone PV
power systems. A few megawatt-size utility-connected
PV installations, made possible by Government
sponsorship and tax incentives, have demonstrated
the technical feasibility and excellent reliability of large
multi-megawatt PV power-generation plants.
Early needs of the space program for electrical
power were satisfied by crystalline silicon solar cells.
During the 1960s and 1970s, it was demonstrated
that PV was a dependable electrical power source for
spacecraft. In this time interval, solar-cell quality and
performance improved and the costs decreased.
However, the costs were still much too high for
widespread use on Earth. There was a need to .
drastically lower the manufacturing costs of solar cells
if they were to be a practical, widely used terrestrial
power source.
Although fuel prices have leveled off, worldwide
interest in photovoltaics continues, especially in
countries where electricity is more expensive than in
the United States. It is anticipated that the field of
photovoltaics, with its tremendous potential, will
regain its place in the sun as fossile fuel energy
sources deplete and new energy technologies
continue to evolve.
In 1975, the U.S. Government initiated a terrestrial
PV research and development (R&D) project with the
ii
Flat-Plate Solar Array
Project Summary
The Flat-Plate Solar Array (FSA) Project, a
Government-sponsored photovoltaics project, was
initiated in January 1975 (previously named the LowCost Silicon Solar Array Project) to stimulate the
development of PV systems for widespread use. Its
goal then was to develop PV modules with 10%
efficiency, a 20-year lifetime, and a selling price of
$0.50 per peak watt of generating capacity (1975
dollars). It was recognized that cost reduction of PV
solar-cell and module manufacturing was the key
achievement needed if PV power systems were to be
economically competitive for large-scale terrestrial
use.
the major limitation. Continuing solar-cell and module
cost reductions are still the key activities. However, it
is now known that, for a utility plant, area-related
costs are significant enough that flat-plate module
efficiencies must be raised to between 13 and 17 %
(more energy generated per fixed-area dollar), and
module life should be extended to 30 years. It is
believed by FSA reseachers that both of these
requirements and the corresponding cost reductions
are possible, but are achievable only with a dedicated
effort. Present FSA activities emphasize high
efficiency, long life, reliability, and low cost for solar
cells, modules, and arrays.
The project was initiated at JPL to meet these
goals through R&D of all phases of flat-plate module
technology, from solar-cell silicon material refinement
through verification of module reliability and
performance. The Project sponsored parallel
technology efforts with periodic progress reviews and
continuing sponsorship of only the most promising
options. A module manufacturing cost-analysis
capability was developed that permitted cost goal
allocations to be made for each module technology,
based upon potential for achievement. Economic
analyses, done as technical progress was achieved,
permitted assessments to be made of each technical
option's potential for meeting the goal and of the
Project's overall progress toward the national goal.
Major Project
Accomplishments
«I
Established the basic technologies for all
aspects of manufacturing and evaluated nonconcentrating crystalline silicon PV modules and
arrays for terrestrial use.
• Established a new low-cost, high-purity siliconfeedstock-material refinement process.
111
Excellent technical progress across the entire
project was accomplished over the years, with
growing interest and participation by the private
sector. More recently, effective energy conservation
practices, a leveling of energy prices, a perception
that energy prices are not increasing endlessly, and a
change in Government emphasis has altered the
picture for photovoltaics. DOE's National
Photovoltaics Program was redirected to longer-range
efforts that the private sector avoids because of
higher risk and longer payoff time. Consistent with
these new directions, FSA has been concentrating its
efforts on overcoming specific critical technological
barriers that inhibit large-scale PV use. These
activities require an in-depth, long-range, integrated
effort that industry cannot reasonably be expected to
undertake alone. This work is being performed by a
team that includes universities, industry (which shares
the cost), and Government laboratories. This team
has worked together successfully for 10 years.
Made significant advances in quality and cost
reduction of:
• Silicon sheet for solar cells.
• Higher efficiency solar cells.
• Manufacturing processes.
111
111
111
An estimate that PV-generated power should cost
about $0.15/kWh in the 1990s to be competitive in
utility central-station generation plants is the basis for
a DOE Five-Year Research Plan. The high cost of PV
generating equipment, especially modules, remains
iii
Established PV module/array engineering/design
and evaluation knowledge and capabilities.
Established PV module encapsulation systems.
Devised manufacturing and life-cycle cost
economic analysis capabilities.
«I
Transferred the above technologies to the
private sector by dynamic interactive activities
including hundreds of contracts, comprehensive
module development and evaluation efforts, 25
Project integration meetings, research forums,
presentations at hundreds of technical meetings,
and advisory efforts to industry on specific
technical problems.
III
Stimulated the establishment of a commercial
PV industry in the United States.
Photovoltaics
Community
Interactions
has been the conduct of 25 Project Integration
Meetings (PIMs) over the 10-year life of the Project.
The PI Ms have been conducted to enable the rapid
exchange of data between all FSA participants,
assess recent progress, gain perspective on trends
and new developments, and to guide the FSA
Project's planning and priority adjustment. The
communication has been open between the Federal
program and the PV community at all levels of detail
regarding problem-solving and choice-making. At
each PIM, contractors presented their achievements
and problems encountered since the previous PIM.
These presentations to their peers and the ensuing
discussions stimulated extensive technical
exchanges and progress. DOE management
participated in every PIM, lending the guidance
needed for credible technology transfer.
Following the OPEC oil embargo of 1973, a
conference of photovoltaics experts was held at
Cherry Hill, New Jersey, to assess the viability of
photovoltaics as a terrestrial electrical energy
source. In response to a 1974 request from the
National Science Foundation, JPL proposed a PV
module R&D effort based upon the Cherry Hill
conference recommendations. This was accepted in
January 1975 and a Project, now called Flat-Plate
Solar Array, was initiated.
Changes in the overall political environment
resulted in the FSA reducing its scope of activities
starting in 1981. The DOE thrust became less
oriented toward the direct development of the U.S.
photovoltaics industry infrastructure and more toward
the research into long-term, high-risk, high-payoff
technical options that industry would not likely
undertake themselves.
The JPL Project quickly established relationships
with university research teams, non-profit
organizations, and commercial firms interested in the
development of low-cost photovoltaics. The
fundamental strategy adopted by JPL was to involve
industry and university teams in a highly interactive
effort along parallel paths of technology
development, and to select the best options. The
effort has been highly goal oriented using a scheme
of cost allocations to guide the progress of the
activities.
Continuing under the new charter, FSA developed
the idea of Research Forums on specific technical
problems of special difficulty wherein a working
group of experts both within and outside the PV
community would convene. Of the many such forums
held to date, everyone has resulted in new ideas
and contractual relationships to help solve the
problem.
By 1978, the MIT-Lincoln Lab; MIT-Energy Lab;
Brookhaven National Lab, the NASA-Lewis Research
Center; Sandia National Lab; a newly formed SERI;
the U.S. Army MERADCOM; and others had become
members of the Federal program under the newly
established U.S. Department of Energy (DOE), which
replaced the Energy Research and Development
Administration (ERDA). The JPLlFSA became very
much involved through many technical interfaces
with all of the DOE Laboratories in order to carry out
the national mandate.
The legacy of the FSA Project consists of
significant measurable technical and economic
achievements. An equally enduring part of the legacy
may well turn out to be the way in which the Project
was able to work both effectively and efficiently with
the PV community to achieve the progress to date.
Perhaps the most important strategy in ensuring
technology transfer has been to directly involve the
recipient photovoltaics community. The single, most
prominent effort in the technology transfer strategy
The FSA organization has been changed
throughout the years to meet the programmatic
needs and has evolved to its present form as shown
below in the Project Organiza.tion Chart.
FLAT·PLATE SOLAR ARRAY
PROJECT
W. T. Callaghan, Manager
D. G. Tustin, Admin. Manager
I
I
I
I
PROCESS
DEVELOPMENT
MATERIALS
AND DEVICES
PROJECT ANALYSIS
AND INTEGRATION
D. 8. Bickler,
Manager
M. H. Leipold,
Manager
RELIABILITY AND
ENGINEERING
SCIENCES
I
A. G. Ross,
Manager
P. A. McGuire.
Manager
I
I
I
I
I
SILICON
MATERIALS
SILICON SHEET
DEVICE RESEARCH
WEB TEAM
A. O. Morrison,
Manager
A. D. Morrison,
Manager
A. H. Kachare,
Manager
R. R. McDonald.
Manager
Iv
Contents
Project Overview . . . . . .
........... 1
Project Technical Thrusts.
.
FSA Project Technologies
2
................... 5
Project Analyses and Integration
.7
Silicon Material
.12
Silicon Sheet .
.23
High-Efficiency Solar Cells
.38
Process Research and Development
.42
Modules and Arrays
.53
Module and Array Evolution
.58
Engineering Sciences and Design
.66
Encapsulation
.70
Reliability Physics
.77
Module Performance and Failure Analysis.
.82
Appendix A: Glossary
.88
.
Appendix B: FSA Project Contractors
Acknowledgment
The progress of the past 10 years in crystalline
silicon PV technology (non-eoncentrating) has been
accomplished only because there has been a strong
motivation to work on a very challenging and
worthwhile activity in which many people believe.
Without the sustaining support of the Federal
Government and cost sharing by private industry,
the excellent achievements across the complete
broad scope of activities would not have progressed
so rapidly. The thousands of people in many diverse
large and small organizations deserve credit for the
excellent work performed individually and as a team.
This report, which summarizes 10 years of FSA
activities, was possible only with the help of many
other members of the FSA Project.
.91
Prepared by the Jet Propulsion Laboratory, California Institute of Technology,
for the U.S. Department of Energy through an agreement with the National
Aeronautics and Space Administration.
The JPL Flat-Plate Solar Array Project is sponsored by the U.S. Department of
Energy and is part of the Photovoltaic Energy Systems Program to initiate a
major effort toward the development of cost-competitive solar arrays.
This report was prepared as an account of work sponsored by an agency of the
United States Government. Neither the United States Government nor any
agency thereof, nor any of their employees, makes any warranty, express or
implied, or assumes any legal liability or responsibility for the accuracy I comp~eteness, or usefulness of any information, apparatus, product, or process
disclosed, or represents "that its use would not infringe privately owned rights.
Reference herein to any specific commercial product, process, or service by trade
name, trademark, manufacturer, or otherwise, does not necessarily constitute or
imply its endorsement, recommendation, or favoring by the United States
Government or any agency thereof. The views and opinions of authors
expressed herein do not necessarily state or reflect those of the United States
Government or any agency thereof.
This publication reports on work done under NASA Task RE-152, Amendment
66, DOE / NASA IAA No. DE-AIOI-76ET20356.
Project Overview
Objectives
• To develop the flat-plate PV array technologies
required for large-scale terrestrial use late in the
1980s and in the 1990s:
• Develop module and array specifications,
develop performance and lifetime criteria, and
assess contemporary module and array
technologies.
• To advance crystalline silicon PV technologies.
• To develop the technologies required to convert
thin-film PV research results into viable module
and array technology.
• To stimulate transfer of knowledge of advanced PV
materials, solar cells, modules, and arrays to the
PV community.
• Continue to fund and cost-share efforts with
industry, universities, and other institutions.
• Continue economic analyses to evaluate
technical options and overall progress.
The definition of a solar cell, module and array is
shown in Figure 1.
Plans
DIRECT SUNLIGHT ON A SlllCON
SOLAR CELL {10 em OR 4 in, DiAl
• To sponsor or cost-share efforts that industry
reasonably would not fund because of high risks
and/or long-time requirements:
• Perform advanced solar cell, module, and array
technology R&D that the private sector can use
to establish economical central station and
rooftop applications, competitive with largesystem power-generation costs of $0.15/kWh in
the 1990s.
• Emphasize efforts to reduce technical barriers
to the development of high-performance,
economical, and long-life solar cells, modules,
and arrays.
• Establish the technologies for a low-cost,
15 % efficient, 30-year-life, crystalline-silicon
flat-plate module.
• In cooperation with SERI, establish the
technology for a low-cost, 12% efficient, longlife, thin-film module.
SOLAR CELL
TRANSPARENT
SEAL
-"-li~iiiiiiii-i~"'--'
METALLIC CONTACTS
TOP AND BOTTOM
ABOUT
1PRODUCES
WATT OF DIRECT
CURRENT ELECTRICAL
POWER 1::::112 volt,
2amporosl
MODULE -
MANY SOLAR CElLS
IN A WEATHERPROOF UNIT
ARRAY -
MANY MODULES ELECTRICALLY
AND PHYSICALLY CONNECTED TOGETHER
Figure 1. Definitions
Project Technical Thrusts
The three flat-plate-array technical thrusts now
required to meet the DOE Program goal of PV power
system generations costs of $O.15/kWh are:
been set for each thrust and for specific contributing
efforts; the contribution and potential contribution of
individual efforts are periodically assessed, as is the
combined contribution toward each thrust. The Project
as a whole contributes to each thrust; the synergistic,
integrated value of that contribution is important, and
is likewise assessed. The goals of each thrust are the
result of years of integrated assessments of user
needs to meet competition and the potential of each
technology to meet the requirements. These efforts
are all interrelated.
• High efficiency.
• Low cost.
• Reliability.
To achieve the overall goal, individual goals have
Ob"jectlves:
Lower Cost
EI Higher Efficiency
EI Higher Reliability
Module/Array
Design
Technology
Encapsulation
Materials &
Processes
EI
~
J
Silicon
Refinement f-Processes
-
Process
Technology
EI Sheet to solar
cells
Material & Cell
Characterization
& Enhancement
l-
AND
EI Cells assembled
into modules
t
t
Silicon Sheet
Formation
EI Ribbon grown to
desired thickness
~
t
t
t
-
Advanced Prototype
Modules
EI JPL specifications
EI Designs by Industry
EI Mutual design review
EI Fabrication by Industry
t
t
Project Analysis and Integration
EI Price allocation guidelines
EI Economic analyses
EI Probability of success
EI Industry assessment
- - - Information Flow, Both Directions
2
t
--
t
Test
EI Performance
EI Durability/Reliability
Failure Analysis
EI JPL tests
EI Applications
Results to Industry
High Efficiency: Crystalline Silicon
Ell
Factors contributing to more efficient solar cells,
modules, and arrays:
ED
ED
Ell
• Translation of thin-film research and the blending
of crystalline silicon PV experience into
commercial manufacturing technologies for thinfilm PV power units is now being accelerated by
industry and by FSA in support of SERI. FSA is
involved in:
High-purity silicon material.
High-quality silicon sheet as grown or
deposited.
ED
Advanced solar cell designs.
ED
Advanced module and array designs.
ED
Advanced cell processes.
ED
Advanced module processes.
ED
ED
Low Cost: Crystalline Silicon
• Factors contributing to low-cost solar cells,
modules, and arrays are:
ED
Understand the influence on efficiency of:
• Silicon growth-related defects including:
ED
• Impurities.
• Structural defects.
ED
• Advanced solar cell designs.
ED
• Advanced cell processing.
ED
ED
ED
Develop advanced silicon sheet growth
capabilities for high quality sheet that is also
inexpensive.
ED
Develop advanced material and solar cell
property measuring techniques and
equipment.
Low-cost, high-purity silicon semiconductor
material.
Rapid growth or deposition of silicon in sheet
form of a quality consistent with cost-efficient
solar cells.
Efficient solar-cell designs that are
inexpensive to fabricate.
High-volume cell processes using inexpensive
materials that have high yields of costeffective solar cells.
Efficient module designs using inexpensive
materials that are easily mass-produced with
high yields of long-life, reliable modules.
• R&D is required to achieve:
Develop advanced solar cell designs and
fabrication processes.
ED
ED
ED
Develop advanced module and array designs.
ED
High Efficiency: Thin-Film Solar Cells
Ell
Developing and verifying module and array
technologies for reliable units.
R&D is required to:
ED
Ell
Developing thin-film cell and module process
appropriate for large-area modules.
Thin-film solar cells consist of thin layers of
semiconductor materials deposited on a
substrate so that the cell can absorb and
convert usable light to electricity in a small
thickness. The potential for depositing cells and
cell interconnections in a continuous production
line holds promise for low-cost thin-film modules
that require small amounts of semiconductor
materials.
ED
ED
Completion of silicon refinement research.
Cost-effective silicon sheet formation resulting
in high-quality, inexpensive sheet.
High-efficiency solar cell designs for costeffective processing and performance.
Mass-production processes with high yields of
high-efficiency, inexpensive solar cells and
modules.
Verification of long-life effectiveness of
module and encapsulation materials.
Low Cost: Thin-Film Solar Cells
Achievement of higher-efficiency cells and
modules is the key to widespread use of thinfilm technology for PV power generation.
• Basic research on materials and devices leading
to higher performance is being sponsored and/or
performed by SERI for DOE, and by the private
sector.
3
Ell
Comprehensive manufacturing technology
efforts have not been accomplished as yet
because material, device, and basic cell and
module designs are still evolving.
ED
A thorough assessment and subsequent efforts
are required when the research results are
definitive.
Reliability: Crystalline Silicon
e
Module and array reliability means consistent high
performance for 30 years.
• Structural failure.
e Factors contributing to high reliability are:
e
Stable, long-life materials.
e
Compatible solar cell and module materials.
e
Appropriate cell and module designs.
e
Suitable cell and module processes.
e
Soiling or dirt accumulation on face of
module.
Design and fabrication flaws.
e Methods of reliability verification are:
e
Development of precise performancemeasurement capabilities.
e
Accelerated testing of materials and modules.
• Module qualification testing.
• High quality of manufacturing workmanship.
e
e
Capability of integration into PV power
systems and installation.
Field testing of modules and arrays.
e R&D is required for:
e Types of reliability or performance degradation
e
are:
Advanced module and array specification
definitions.
e
Infant mortality.
• Material aging and stability research.
e
Wear-out or aging.
e
Advanced mOdule and array design
technology research.
e
Random failures.
e
e
Environment-induced soiling (weathervariable).
Solar-cell, module, array, and PV system
compatibilities.
e
Advanced module design verification.
e Causes of degradation are:
• Correlation of accelerated testing with field
testing and with operational results.
• Cell cracking.
e
Corrosion.
Reliability: Thin-Film Solar Cells
• Cell-interconnect fatigue.
• Module material aging.
e
Module material fatigue.
e
Insulation failure.
e Unknown, evaluation just beginning.
e Most of the crystalline-silicon technology, noted
above, will apply to thin-film hardware, but
probably with unique variations.
4
FS
Project Technologies
SILICON MATERIAL
SILICON SHEET
• QUARTZ (SI02lIS REFINED BY REMOVING 02 AND IMPURITIES
• THE PRODUCT IS VERY HIGH PURITY SILICON
• SILICON IS GROWN WITH SPECIFIC ELECTRICAL AND
CRYSTALLOGRAPHIC CHARACTERISTICS
• INGOTS ARE SLICED INTO WAFERS
• RIBBONS ARE GROWN TO DESIRED THICKNESS
RIBBONS
INGOT
WAFER
NEW SILICON REFINEMENT PROCESSES
ACCOMPLISHMENTS
• LARGE·SCALE SILANE PRODUCTION PLANT BY UCC IN OPERATION
• PRACTICAL CONVERSION OF SILANE TO SILICON IN FBR DEMONSTRATED
• OTHER PROCESS AND PROCESS IMPROVEMENTS ADOPTED BY INDUSTRY
ACCOMPLISHMENTS
•
•
•
•
RESEARCH REQUIRED
• ESTABLISH PURITY OF SILICON FROM UCC PROCESS
• VERIFY ECONOMICS OF PROCESS
NEW SEMICONTINUOUS, LARGER CZ INGOTS GROWN
NEW SHAPED·INGOT TECHNOLOGIES EVOLVED
FASTER, MORE EFFICIENT SAWING DEVELOPED
GROWTH OF WIDE, FLAT·RIBBON ACHIEVED
RESEARCH REQUIRED
• INVESTIGATE CRYSTALLIZATION RATE LIMITATIONS
AND INFLUENCE OF PRIMARY GROWTH PARAMETERS
• CONTINUE CHARACTERIZATION OF SILICON SHEET
MATERIAL VERSUS GROWTH RATES
• CONTINUE BASIC INVESTIGATION OF HIGH·QUALlTY,
RAPID CRYSTAL GROWTH; ESPECIALLY RIBBON GROWTH
HIGH·EFFICIENCY SOLAR CELLS
SOLAR CELL EFFICIENCY IS A FUNCTION OF:
• SILICON SHEET QUALITY
• CELL DESIGN
• CELL FABRICATION PROCESSES
UNIFORMITY AND REPEATABILITY OF ABOVE
PROBABLE MAXIMUM ACHIEVABLE_
25
eft
20
>()
LABORATORY
CELL AND MODULE PROCESS RESEARCH AND DEVELOPMENT
• SOLAR CELLS ARE FABRICATED FROM SILICON SHEET
CELLS ARE ELECTRICALLY INTERCONNECTED AND ENCAPSULATED
IN A SUPPORTING STRUCTURE
CELLS_-------..._~'---------l
Z
~
15
u::
u.
UJ
...J
...J
UJ
()
o L :-:::---------,-:Y=-EA,.-R::---------,-19=-=a,.,J5
197 5
ACCOMPLISHMENTS
• KEY BARRIERS TO HIGHER·EFFICIENCY
SOLAR CELLS IDENTIFIED
• HIGH·EFFICIENCY SOLAR CELLS MODELED
AND ANALYZED
• NEW SOLAR CELL CHARACTERIZATION
TECHNIQUES DEVELOPED
ACCOMPLISHMENTS
• IMPROVED CELL SURFACE TREATMENTS, JUNCTION FORMATION, AND METALLIZATION
HAVE RESULTED IN LESS EXPENSIVE YET MORE EFFICIENT SOLAR CELLS
• SIMPLIFIED FABRICATION SEQUENCES SUITABLE FOR AUTOMATION DEVELOPED
• PROTOTYPE MASS·PRODUCTION EQUIPMENT DEVELOPED
RESEARCH REQUIRED
RESEARCH REQUIRED
• CONTINUE INVESTIGATION OF LOSS MECHANISMS
WITHIN SOLAR CELLS
• CONTINUE DEVELOPMENT OF SOLAR CELL CHARACTERIZATION
ANALYSIS AND MEASUREMENT TECHNIQUES
CONTINUE RESEARCH CELL DEVELOPMENT AND TESTING
• INVESTIGATE HIGH·EFFICIENCY SOLAR CELL PROCESSING
• INVESTIGATE ELECTRICALLY CONDUCTIVE SILICIDES, CORROSIVE
EFFECTS ON METALLIZATION, AND THE EFFECTS ON JUNCTION FORMATION
• CONTINUE INVESTIGATING CRITICAL PROBLEMS OF ION IMPLANTATION,
METALLIZATION, INTERCONNECTIONS, AND MODULE ASSEMBLY
5
MODULE/ARRAY ENGINEERING SCIENCES AND DESIGN
ENCAPSULATION
DESIGN AND TEST METHODS HAVE BEEN DEVELOPED AND USED TO IMPROVE
THE PERFORANCE, ENVIRONMENTAL DURABILITY, AND SAFETY OF
PHOTOVOLTAIC MODULES AND ARRAYS
SOLAR CELLS AND INTERCONNECTS ARE PROTECTED FROM
THE ENVIRONMENTS
TRANSPARENT COVER - -
FIRE TEST
OF MODULES
SOLAR CELLS AND
ELECTRICAL
INTERCONNECTIONS
BACK COVER
ARRAY
STRUCTURE
TEST
----~• • • • • • • •II,..
ACCOMPLISHMENTS
• EXPERIENCE INDICATES 30·YEAR MODULE LIFE IS ACHIEVABLE
• SELECTED MATERIALS AND PROCESSES EXPECTED TO MEET GOALS
AND AUTOMATED PROCESSING NEEDS
• MATERIALS INDUSTRY RESPONDING TO NEEDS AS DEFINED BY FSA
RESEARCH REQUIRED
• CONTINUE INVESTIGATION OF ENCAPSULATION MATERIALS TO
ACHIEVE LONG·TERM PHOTOTHERMAL STABILITY
• CONTINUE DEVELOPMENT AND USE OF ACCELERATED AND LONG·LIFE,
DURABILlTY·TESTING TECHNIQUES
• CONTINUE INVESTIGATION OF ENCAPSULANT, METALLIZATION, AND
OTHER INTERFACIAL BONDING STABILITY AND CORROSION CRITERIA
AND MECHANISMS
ACCOMPLISHMENTS AND ONGOING ACTIVITIES
• MODULE AND ARRAY DESIGN REQUIREMENTS GENERATED FOR FUTURE
LARGE·SCALE APPLICATIONS
• IMPROVED DESIGN CONCEPTS, ANALYSIS, AND TEST METHODS
AVAILABLE AND UNDER DEVELOPMENT (CIRCUIT DESIGN, SAFETY,
STRUCTURES, RELIABILITY, ETC.)
• PERFORMANCE EVALUATION METHODS AND DESIGN AND TEST
SPECIFICATIONS DEVELOPED THROUGH INTERACTION WITH INDUSTRY
RESEARCH REQUIRED
• PERFORM ABOVE FUNCTIONS FOR MODULES
MADE FROM THIN·FILM SOLAR CELLS
RELIABILITY PHYSICS
POWER GENERATION WITHOUT DEGRADATION REQUIRES:
MODULE PERFORMANCE AND FAILURE ANALYSES
• STABLE, LONG·L1FE MATERIALS
• GOOD, COMPATIBLE CELL, MODULE, ARRAY AND PV SYSTEM MATERIALS
AND DESIGN
• GOOD FABRICATION AND INSTALLATION WORKMANSHIP
• PHOTOTYPE ADVANCED MODULES HAVE BEEN DESIGNED AND MANUFACTURED
BY INDUSTRY TO JPL SPECIFICATIONS
• MODULE ELECTRICAL PERFORMANCE AND ENVIRONMENTAL DURABILITY
ASSESSMENT BY JPL
(J)
ex:
«
UJ
>-
1975
1985
ACCOMPLISHMENTS
• DEVELOPED ANALYSIS TECHNIQUES FOR EVALUATING PARAMETRIC
INFLUENCES ON RELIABILITY
• DEVELOPED MANY TEST TECHNIQUES AND EQUIPMENT FOR
EVALUATING MODULE RELIABILITY
• MODULE LIFE·TIMES HAVE INCREASED DRAMATICALLY
RESEARCH REQUIRED
• CORRELATE FUTURE FIELD OPERATIONAL RESULTS WITH EXISTING
DATA
• EVALUATE AND IMPROVE THIN·FILM MODULE RELIABILITY
PROGRESS IN MODULE PERFORMANCE
• MODULE PERFORMANCE AND DURABILITY IMPROVEMENT WITH EACH
NEW PROCUREMENT
• LATEST MODULES (5th PROCUREMENT) INCORPORATE INNOVATIVE
DESIGN FEATURES
• CONTINUING CHARACTERIZATION AND DIAGNOSTIC ANALYSES OF
MODULE DEFICIENCIES/FAILURES ARE TRANSFERRED TO INDUSTRY
PROJECT ANALYSIS AND INTEGRATION
• DEVELOPED PV ANALYSIS METHODS, ESPECIALLY FOR
MANUFACTURING COSTS
• ESTABLISHED GOALS INTERNAL TO FSA AND MEASURED TECHNICAL
PROGRESS BY USE OF ANALYSIS TECHNIQUES
6
Project Analysis
and Integration
the original Project structure to address the problem
of communication and to help guide the Project
technical developments toward the National PV
Program's economic goals.
Project Activities
Those responsible for initiating and organizing the
FSA Project recognized the need for coordination
. within the Project and with private industry to
achieve the objectives established for the Project. To
that end, the Project Analysis and Integration (PA&I)
effort has served as a channel of communication for
Project activities, private industry, and the U.S.
Department of Energy.
In 1975, analytical capabilities were needed to
evaluate competing technologies and establish
realistic guidelines for their development. Initially,
cost goals were allocated to silicon material, silicon
sheet, encapsulation materials, cell fabrication
processes, and module assembly. As economic
analysis capabilities were developed, thereby
allowing Project-wide technology assessments to be
made, these goal allocations were revised. The
technical objectives that were established were used
to determine the division of resources throughout the
Project. As time passed, the completeness and
sophistication of the analysis capabilities matured.
PA&I performs planning and integration activities
to coordinate the various R&D tasks. Goals are
established for Project R&D activities that reflect the
objectives of the National Program. Analytical
support is given to the Project Office, providing a
channel for the integration of individual task work into
the overall Project. Integration activities have also
supported the coordination of research activities
between the Project and other research laboratories.
The assessing of technology status and the
tracking of progress has been one of the principal
PA&I activities throughout the duration of the FSA
Project. It has established a set of standards by
which progress can be measured. The standards are
part of an overall methodology developed by PA&I so
that research findings could be reported in a
common format. The methodologies are being used
to track and assess progress within the Project so as
to derive the maximum technical and economic
benefits from research and development work.
Technical reports and research activity assessments
are an integral part of the overall Project effort.
PA&I performs analytical studies to measure the
progress toward accomplishing Project goals and to
determine which R&D tasks are the most likely to
result in significant performance and cost
improvements. Guidelines for module production
costs are established by PA&I which are consistent
with the National Program goals. Performance and
cost criteria are set for separate module elements,
creating objectives for the individual R&D tasks.
Sensitivity studies are performed to see how
technical advances in the Project and changes in the
economic environment affect Project objectives.
The FSA Project participation in the early efforts of
DOE to establish a National PV Program Plan was
coordinated and implemented by PAM It authored
several major sections of the first DOE Program
Plan. The planning and organization of the
Photovoltaic Technology Development and
Applications Lead Center at JPL in 1978 was a major
PA&I activity at that time.
Background
FSA research and development has encompassed
all phases of flat-plate module technology, from the
basic feedstock materials for the photovoltaic
manufacturing processes, through vehfication of
module reliability and performance in the field. When
the Project was organized in 1975, it was recognized
that this broad spectrum of activities would require
the Project to be somewhat unique at JPL because
of the number of industrial organizations, Government
laboratories, academic institutions, and other entities,
such as utilities, with which JPL would have to
interact. It was apparent at the outset that
communication between JPL and the other
participating organizations would be an important
factor in the success of the effort. In addition, the
Project was one of the first, and certainly the largest
at JPL, for which technical performance was not the
ultimate goal but only a contributing factor to the
principal measure of success,economic
performance. Thus, it was also apparent that
analytical methods would be needed to translate
technical performance into future economic
performance. The PA&I Task was formed as part of
Today's Status
PA&I provides ongoing support to the Project in
the form of up-to-date technology assessments.
ManUfacturing technology descriptions are revised to
reflect new technical developments,and new
manufacturing cost estimates are made using the
SAMICS .(Solar Array Manufacturing Industry Costing
Standards) methodology. Examples of this analysis
capability are the 1985 production cost estimate
updates for Czochralski (Cz) and dendritic web silicon
solar cell modules.
Module cost estimates were made for state-of-theart ,Cz silicon solar cell technology, assuming the
construction of a 25-MW-per-year production facility
and current material costs. Assuming 3 shifts per
day, full-time operation and module efficiencies of
13.5%, production prices were estimated at $1.47 per
7
peak watt (1985) dollars. This represents significant
progress over early assessments for this technology.
The SAMICS methodology is a complete procedure
for evaluating the cost of manufacturing PV solar
arrays. The SAMICS methodology includes:
Module cost estimates were made for state-of-theart, dendritic web silicon solar cell technology,
assuming three shifts per day, full-time operation at a
25-MW-per-year production rate and current material
costs. Assuming module efficiencies of 13.7 %, production prices were estimated at $1 .48/Wp or more.
This is based upon a web growth rate of 10 cm 2/min
or less (today's best growth rate).
• Computer programs designed for user
implementation on personal computers (SAMIS
PC, SAMPEG, IPEG-4).
• Formats for the preparation of complete
manufacturing process descriptions.
• A detailed cost account catalog which can
easily be modified to meet user requirements.
The dendritic web solar cell technology, while still
under active development, promises to be a most
attractive photovoltaic option. The principal
impediment to economic feasibility of dendritic web
has been the low crystal growth rates. If the
technology development presently in progress is
successful in increasing the dendritic web growth
rates to 30 cm 2/min, then the production price can
be reduced to $0.86/W p (1985 dollars). If modules at
this price are placed in a PV electrical· generation
plant, electricity could be produced at $0.17/kWh,
very close to the DOE Program goal of $0.15/kWh.
• A standard set of financial assumptions making
direct technology comparisons possible.
• A report format that provides detailed value
added estimates for all process steps.
• The option to vary any of the technical,
financial, or overhead parameters.
A series of competitive procurements, designated
Block I through Block V, was initiated by FSA to
encourage the adoption of new technology and
stimulate reduction in the cost of photovoltaic
module production. SAMICS has made it possible to
assess the technical progress made. Detailed
process descriptions reflecting the latest technology
developments were converted into manufacturing
cost estimates under commercial production
conditions. These were then compared with previous
block purchases and the cost goals of the National
Program.
Accomplishments
PA&I supports a systems analysis capability for
evaluating developments within the Project and
assessing progress. The present status of this
capability and its accomplishments are best
summarized in terms of the methodologies or
analytical tools maintained for Project analysis and
integration efforts. These analytical tools have
provided a consistent and reliable framework for
measuring developments throughout the Project.
SIMRAND
SAMICS
Analytical tools. have been developed which were
designed specifically for support of program
management activities. The FSA Project must
continually make decisions regarding support levels
fot various projects. These decisions are necessarily
based on information that is quite uncertain in
nature. The SIMRAND (Simulation of Research and
Development Projects) model was developed as a
tool that allows this uncertainty to be directly
incorporated into the information used by decisionmakers.
One of the most important analytical tools
developed and supported through PA&I is the Solar
Array Manufacturing Industry Costing Standards
(SAMICS). SAMICS has proven itself as a credible
costing methodology for an emerging technology.
The SAMICS methodology provides a standardized
way for comparing cost estimates of many
competing technical options.
SAM ICS has been used repeatedly to assess the
cost of photovoltaic module production at output
levels representative of commercial ventures using
state-of-the-art technology. SAMICS has also been
used for projections of what might reasonably be
expected at some future date for the technology
given successful completion of research activities.
These snapshots of technology development in
progress have provided FSA valuable insights into
the likelihood of achieving desired reductions in the
cost of photovoltaic module production.
The SIMRAND model assesses the probability that
a combination of R&D activities will meet a given set
of price goals. The descriptions developed by
SIMRAND for each technology reflect the uncertain
nature of the database used. Technologies can then
be compared to see which has the highest probability
of reaching Project goals. The decision process is
modeled by means of alternative networks that
represent the feasible technologies being considered.
SIMRAND performs repeated simulations to
8
ECONOMIC ANALYSES
SAMICS - THE SOLAR ARRAY
MANUFACTURING INDUSTRY COSTING STANDARDS
SUM OF
DIRECTWORI<
STATION REQTS
0.0175 MAINT. MECH.
O.026kWhlmln
SAMICS OUTPUTS
DESCRIPTION OF
SIMULATED FACTORY
(INCLUDING FINANCES)
JPL-+FORMAT A
CELLSTANG
400-cufllmlnFUMES
1.191E-4IblmInSOLDER
5.24cull/mln
COMPRESSED AIR
PROOUCT:
STRINGS/min
24 CONNECTORS/min
REQUIRED PRODUCT:
TESTe
99.9% VIELD
ISTR1NG/9CELlS
A FACTORY lOR SEQUENCE
OF FACTORIESllS BEING
BUILT, STAFFED, AND
OPERATED IN THE COMPUTER; THE SIMULAliON
CONCENTRATES ON THE
INDIRECT REQUIREMENTS
(OVERHEADl AND THE
OTHER ECONOMIC ASPECTS
OF THE COMPANY (OR
COMPANIES
INTERNAL DATA BASE
9
The SAMICS and SIMRAND models and the
Allocation Guidelines have been applied individually
to specific technical problems and as an integrated
set when longer range management decisions have
been required. Technical options have been
evaluated by applying the manufacturing cost
simulations to processing sequences and comparing
the results with the cost-goal allocations. Manufacturing cost point estimates generated by SAMICS
have been combined by using SIMRAND with the
uncertainties associated with those estimates to form
probability distributions of technical performance
and/or cost. The inclusion of uncertainty has allowed
Project management insight into the risks associated
with the technical options being considered.
determine the likely distribution of outcomes for each
technology. The network reflecting the various
technology options is continually reevaluated during
the simulation to find the optimum subset of research
tasks.
SIMRAND is a management decision methodology
that incorporates information uncertainty. The
SIMRAND model (Figure 2) includes:
• The capability of evaluating relatively complex
networks of technical options representing the
feasible choices.
• The ability to replace process parameters with
probability distributions reflecting information
uncertainty.
LCP
Assessment of a PV systems capability to
compete with conventional energy resources requires
simulation of its performance over its lifetime and
recognition of regional variations in solar energy
resources. The Lifetime Cost and Performance (LCP)
model was developed and is now used to simulate
the energy output of a PV power plant over its useful
life. LCP relates the effect of hourly weather conditions, system design and component efficiencies,
electrical design, long run effects of exposure, and
alternative operation and maintenance strategies to
system cost, performance, and value (Figure 3). Fixed
array, one-axis tracking and two-axis tracking systems
can be evaluated at different locations throughout the
country.
• The option to include decision-maker's
preferences regarding risk into the decision
process.
• Computer program available for implementation
on a personal computer.
STEP 1
SILICON
PURIFICATION
I
STEPZ
CRYSTAlUZATlON
I
STEP 3
STEP 4
STEP 5
SAW1N~
CEll
MOOUlE
LCP simulates the performance of PV systems
using actual solar resource measurements as
recorded in weather data tapes. The LCP model
includes:
•
The ability to use insolation data, temperature
measurements, and precipitation and dirt
accumulation measurements to model location
specific performance.
•
The ability to evaluate the effects of balance of
system component failures and electrical
mismatches within modules.
•
A computer model that can be implemented on
a personal computer.
Figure 2. Module Production Network
A joint JPLlSERI silicon sheet material study
successfully used SIMRAND to incorporate
uncertainty about potential technical progress and
material costs into the outlook for each silicon sheet
technology. The product was a set of distributions
indicating the possible range of outcomes for each
research activity. The study results were instrumental
in redirecting the technology mix pursued by the
National PV Program.
LCP, which was used to assist Acurex Corp. in
making economic assessments of alternative designs
10
for the first phase of the Sacramento Municipal Utility
Districts proposed 100-MW PV installation. The performance of each design alternative was simulated over
its lifetime using weather data characteristic of the
location and the parameters of the PV array system.
from a variety of causes have been assessed.
PV ARRA Y can model the effect of specific solar cell
failure mechanisms on system performance and correct for various failed module replacement strategies.
PVARRAY simulates the power output of a PV
array over its lifetime. The PVARRAY model includes:
JULY 1
PVARRAY
The PVARRAY model simulates array performance
over time. Using PVARRAY, the power profiles of
modules representing various state-of-the-art and
advanced designs have been obtained and the
effects of the current-voltage (I-V) changes resulting
.....
1986
I
,,,
..
..
;,
I
""
I
,.,.
I
I
"u,
""
1996
The ability to compare performance of different
module designs.
•
The capacity to consider alternative series
parallel wiring schemes.
•
The ability to evaluate alternative replacement
strategies for different cell failure rates and
diode placements.
•
Computer program that can be implemented on
a personal computer.
The combination of PVARRAY, SAMICS, and LCP
has been used at JPL to estimate the net present
value of energy from a photovoltaic system over its
useful lifetime. Manufacturing costs for the module
production sequence have been estimated using
SAMICS, and the life-cycle cost and performance
have been simulated using the LCP and PVARRAY
models. The three programs have allowed PA&I to
calculate the cost of different module designs and
determine the time-d~pendent economic impact of
design changes by simulating array performance and
life-cycle cost (Figure 4).
Figure 3. Time-of-Day Output and Customer Load
Profile
0
•
....., , ....• •
I
I
I
100._ tMi
[
2006
I
I
I
I
n"
'0"
'01"
' - -n'..
'OIl
2016
Figure 4. Cell Degradation and Lifetime Performance for Three Different Module Designs
11
than one part per million, which possesses
semiconductor properties. In 1975, the available
material of this quality was produced by the Siemens
process and was referred to as semiconductor-grade
silicon. The price then was greater than $50/kgand
was very sensitive to the supply-and-demand
fluctuations. It was known that solar cells could be
made from a less-pure silicon. How much less pure,
with which impurities, and how much of each impurity
could be tolerated, was unknown, nor was such a
material commercially available.
Silicon Material
Semiconductor-grade silicon is the material used to
fabricate crystalline-silicon solar cells, most solid-state
devices, and most integrated circuits. It is very highly
purified, with tight specifications limiting the
concentrations of specific impurities. The
manufacturing yields during the production of solidstate devices and the performance of the devices are
directly related to the quality of the silicon.
Semiconductor-grade silicon, as produced in the
Siemens reactor for the semiconductor industry, has
been and continues to be used by the PV industry for
the production of the great majority of the solar cells.
The cost of the silicon is a minor part of the cost of
producing semiconductor devices, but it has been and
still is a significant fraction of the cost of manufacturing
solar cells. Integrated circuits are very small and many
can be fabricated on one silicon wafer, but only one
solar cell is fabricated from a wafer. The small market
provided by solar-cell users has not justified industry's
development of an inexpensive high-volume silicon
production technique that would result in low-cost
silicon material. Considerable FSA funding has been
devoted toward advancing silicon purification
technology and processes.
Project Activities:
1975T01981
The status of the Siemens process for the
production of semiconductor grade silicon was:
Advantages
High purity (more than adequate at that time)
Acceptable form
Proven process
Potential for improvements:
The objective of the silicon refinement effort is to
overcome the technical Qarriers to low-cost silicon
purification in order to achieve a process capable of
producing high-quality, low-cost silicon (appropriate for
photovoltaic use), at a price consistent with FSA
objectives.
Decreased energy use
Increased production rate
Disadvantages
The goals are to establish silicon material processes
appropriate to:
High-energy consumption process
• Purity adequate for efficient crystalline-silicon
solar cells that meet the requirements of the FSA
Project.
Low conversion rate and yield
Large amount of waste by-product
• Market price of $19/kg (1985 dollars).
High labor costs
• High-volume production plants «1000 MT/
year).
High product cost
A plan was formulated within FSA to develop a
silicon refinement process to meet PV sheet and solar
cell requirements. The plan was to develop low-cost
semiconductor-grade silicon processes; to develop a
solar-grade silicon refinement process (a less-pure, but
unspecified material that should be even less
expensive); and to determine the effects of various
impurities and their concentrations on silicon
refinement processes and on solar cell performance.
The plan consisted of three process development
phases for both the semiconductor-grade and the solargrade process (Table 1). The intent was to establish
the practicality of the processes by production, energy
consumption,and economic criteria. To classify
candidate processes and define the process
specifications, information was required about the
effects of specific impurities on silicon and on solar cell
Background
Silicon, the second most abundant element in the
Earth's crust, occurs naturally primarily as silicon
dioxide and as silicates: sand, quartz, etc. Large
quantities of quartzite, a crystalline form of silicon
dioxide, are reduced in arc furnaces to metallurgicalgrade silicon (mg-Si) for use by man'1 industries. It sells
for $1 to $2/kg and is about 98 to 99% silicon. A
much purer silicon is required to fabricate
semiconductor devices and solar cells. A small portion
(a few thousand metric tons) of mg-Si is used annually
in processes to produce semiconductor-grade silicon.
Most solar cells have been and still are fabricated
from silicon with a total impurity concentration less
12
Table 1. Planned Process Research Phases
Phase
Laboratory-Scale
Experiments
Purpose
Demonstrate process feasibility
Determine reaction kinetics and yields
Identify reactor-material problems
Obtain silicon deposition rates of 0.01 to 1 kglh
Obtain product purity information
Process Development
Units (PDUs)
Demonstrate practicality of reactors and critical process steps
Obtain reactor scale-up information
Characterize critical process components
Establish suitability of reactor materials
Determine process capability by chemical engineering analyses using acquired
data and extended process design
Perform preliminary economic analyses
Experimental Process
System Development
Units (EPSDUs)
Establish technical readiness of integrated processes
Design and build pilot plant EPSDUs with silicon deposition rates of up to
100 MT/year
Obtain operating data to characterize complete process including steady-state
operation
Obtain data for optimization of design and operation of a full-scale production
plant
Confirm product purity at steady-state conditions
Confirm economic analyses
performance. These purity requirements directly
affected the process designs and, thus, the economics
of the candidate processes. The impurity studies
consisted of two phases: the measurements of the
effects on the final product of impurities that are found
in mg-Si or that might be introduced during a process
step, and the determination of the interrelationships of
impurities and the processing steps and their combined
effects on solar cell performance.
parameters based upon data from laboratory-scale
experiments. After an investigation lasting 1 to 3 years,
each of the promising specific processes was then
evaluated for its practicality.
Scaled-up reactors were used for chemical
engineering studies to derive data for mass and energy
balances, process flows, kinetics, mass transfer,
temperature and pressure effects, and operating
controls. This information was used in analyses and
extrapolations that resulted in improved process
concepts and designs. The processes being examined
in June 1977 and a number of supporting stUdies are
shown in Table 2. By this time some process
developments had progressed rapidly.
The feasibility of 11 potentially low-eost processes
were investigated. Their technical feasibilities were
assessed by initially studying basic chemical reactions
and general chemical engineering designs, which led to
an analysis of the chemistry and chemical engineering
13
Table 2. Silicon Material Refinement Contractors in 1977
Technology
Contractor
Semiconductor-Grade Production Processes
AeroChem Research Laboratories, Princeton,
New Jersey
Silicon halide-alkali metal flames
Battelle Memorial Institute, Columbus,Ohio
SiC 14 reduction by Zn using FBR technology
Union Carbide, Sistersville; West Virginia
Pyrolysis of SiH 4 derived from redistribution of
SiHC 13 (With a feed of metallurgical grade silicon)
Motorola, Phoenix, Arizona
Chemical vapor transport and pyrolysis of SiF2
generated from SiF 4 reaction with metallurgical grade
silicon
Solar-Cell-Grade Specifications
Northrop Research, Hawthorne, California
Lifetime and diffusion length measurements
C. T. Sah Associates, Urbana, Illinois
Model and studies of effects of impurities on solar
cell properties
Spectrolab, Inc., Sylmar, California
Solar cell fabrication and analysis
Westinghouse Electric, Pittsburgh, Pennsylvania
Investigation of effects of impurities on solar cell
performance
Monsanto Research Corp., St. Louis, Missouri
Investigation of effects of impurities on solar cell
performance
Solar-Cell-Grade Production Processes
AeroChem Research Laboratories, Princeton,
New Jersey
Use of a nonequilibrium plasma jet
Dow Corning, Hemlock, Michigan
Arc furnace process using purer source materials
Stanford Research Institute, Menlo Park, California
Na reduction of SiF4
Schumacher, Oceanside, California
Reduction of SiHBr3 in an FBR
Texas Instruments, Dallas, Texas
Carbothermic reduction of Si0 2 in a plasma
Westinghouse Electric, Pittsburgh, Pennsylvania
Plasma-are-heater reduction of SiC 14 with Na
Commercial Potential of Processes
Lamar University, Beaumont, Texas
Evaluations of relative commercial potentials of
silicon-production processes developed under the
Silicon Material Task
14
After analytical and experimental studies indicated
that a process was feasible and potentially economical,
and the practicality of its reactors and critical steps
were demonstrated, a comprehensive economic
analysis was made. Professor Carl Yaws, of Lamar
University, and JPL personnel made these economic
analyses which not only calculated the costs to
produce silicon, but also made comparative
assessments of the progress of each process. The
most technically and economically promising
processes were then carefully reviewed before they
were deemed ready for design and development of a
pilot plant. The verification of a process in a pilot plant
was necessary to validate silicon refinement
economics and product purity, because of limited
capabilities in scaling-up chemical processes by
analytical means alone.
In support of these efforts, a basic investigation of
the technology of a high-performance silane fluidizedbed reactor (FBR) has been continued at JPL; research
to characterize the aerosol phenomena involved in the
growth of particles formed by the thermal
decomposition of silane is being performed at the
California Institute of Technology, and an engineering
development of the silane-to-silicon FBR is being
performed by Union Carbide Corporation. Continuing
studies within FSA of the influence of impurities in
silicon on solar cell performance are described here
under the heading, "High Efficiency Solar Cells."
Today's Status
• A silane process has been developed by the
Union Carbide Corp. that seems to be capable of
meeting FSA goals, including price.
Purification processes involving deposition of the
silicon from silane and dichlorosilane were
emphasized. Because these two substances can be
purified relatively easily and, because of their high
reactivity, they can be more easily and completely
decomposed or reduced to form silicon than can
trichlorosilane, which is used in the conventional
Siemens process. Research on other processes that
offered promise for making a less-pure solar-cell-grade
silicon by refinement of metallurgical-grade silicon were
continued because of the potential for even lower-cost
silicon.
• A pilot plant (1 OO-MT/year capacity) has
demonstrated the following (see Figure 5):
• The silane portion of the Union Carbide Corp.
silane process has met FSA goals.
• The silane produced in the first section of the
Union Carbide Corp. process is very high
purity and will be a low-cost material suitable
for amorphous-silicon solar cells and/or other
products when and if they become largequantity products.
A comprehensive study of the effects of impurities
and their concentrations on semiconductor-grade
silicon and crystalline-solar-cell performance was
carried out by a number of organizations, with the
major focus of this effort at Westinghouse Electric
Corp. A functional empirical model was formulated
based upon the relationship that impurity atoms
primarily decrease minority carrier recombination
lifetime, reduce the short-circuit current of a cell, and
essentially act independently. Experimental data
confirmed this and provided quantitative relationships
of the effects of specific elements on cell efficiency.
• The R&D results of the silane-to-silicon
deposition FBRs of the process are good.
Research at JPL, which established the
fundamental technology for FBR silane-tosilicon pyrolysis, will end in 1985. The FBR
development at Union Carbide Corp., which
has been sponsored by DOE, will continue
with Union Carbide Corp. funds.
• Silane process production plants (completely
funded by Union Carbide Corp.) include:
• A 1200-MT/year plant which started operation
early in 1985 converting mg-Si to silane, based
upon DOE-sponsored technology. Komatsu
chemical-vapor-deposition (CVD) reactors
convert silane into high purity silicon at a price
that is expected to be less than that of
Siemens-produced silicon. It is almost at full
capacity now (see Figure 6).
Project Activities: 1981 to Present
In 1981, the funds available were reduced
considerably, which resulted in many activities being
dropped and support being given only to the most
promising, advanced efforts. At about the same time,
the then newest PV system cost assessments indicated
that higher-efficiency solar cells and modules were
necessary to offset the area costs of an array. Analysis
indicated that the power generation cost reductions
possible by the use of higher-efficiency solar cells were
greater than cost reductions possible by use of solarcell-grade silicon. Cell efficiency is directly related to
silicon purity. Therefore, solar-cell-grade silicon seemed
to be impractical economically and efforts were
reduced. The Union Carbide Corporation silane
process and the Hemlock Semiconductor Corp.
dichlorosilane process developments were continued.
• A second 1200-MT/year plant is under
construction with completion scheduled for
1987.
• A third plant of 3000-MT/year capacity is in
design, with completion scheduled late in
1988. Use of a FBR is under investigation.
• The plants listed above will significantly increase
world production of silicon. World capacity to
produce semiconductor-grade silicon was about
6000 MT/year at the end of 1984.
15
Figure 5. A 1OO-MT Silane Process Pilot Plant, Washougal, Washington
Figure 6. A 1200-MT Silane Process Production Plant, Moses Lake, Washington
16
Significant Accomplishments
process of Hemlock Semiconductor Corp. was
successfully carried through the demonstration of the
practicality of the reactor and the critical process steps.
A PDU was designed, constructed, and tested. This
PDU was used to investigate the making of
dichlorosilane from trichlorosilane. A pilot plant was not
sponsored by DOE/JPL because process economics
indicated a product price less than for the conventional
Siemens process, but greater than that of the silane
process. This technology is available for use in a
production plant and Hemlock may have assimilated it
into their ongoing production of silicon, although
Hemlock has never divulged its commercialization
plans to JPL.
The silane process developed at Union Carbide
Corp. under an FSA contract seems to have the
capability to meet the FSA goal of semiconductorgrade silicon at $19/kg (1985 dollars). This process
comprises two parts: the preparation of pure silane
(SiH 4) from metallurgical-grade silicon, and the
deposition of high-purity silicon from the silane using a
FBR. Operation of a 100-MT/year pilot plant has
verified the technical and economical practicality of the
silane portion of the process. The rate of production
and the purity of the silane from the Union Carbide
Corp. pilot plant are consistently high. A silane sample
tested by Brookhaven National Laboratory was the
purest silane it had ever tested. Good research and
development results for low-eost pyrolysis of silane to
silicon are coming from the collaborative fluidized-bedreactor efforts at Union Carbide Corp. and JPL. The
Union Carbide Corp. effort is determining the steadystate operational conditions for the design of a
production prototype reactor for silane feed
concentrations of 20 to 25 %. For more cost-effective
pyrolysis, JPL research is being performed at silane
concentrations up to 100 % with emphasis on 40 to
60% concentrations. The problems of reactor clogging,
bed slugging, and excessive production of silicon dust
have been solved along with raising the silane feed
limits from 2 to 100% concentration. The Union Carbide Corp. FBR has successfully run for 66 h at 20%
silane. Unresolved problems are suitable silicon seed
particle material, prevention of contamination, product
purity confirmation, and verification of production
economics.
Numerous FSA studies also contributed to
advancements in trichlorosilane and modified Siemensreactor technologies. A thorough examination of the
many silicon processes has resulted in significant
improvement in the understanding of many other
silicon refinement technologies. The solar-eell-grade
silicon efforts were stopped because of (1) large
budget reductions, (2) the advanced state of the Union
Carbide Corp. silane process, (3) improved
understanding of the need for higher-purity silicon to
obtain high cell efficiency, and (4) the economic value
of higher-efficiency solar cells.
A good fundamental understanding of the effects of
impurities in silicon upon the properties of silicon
material and upon the performance of solar cells has
been determined by extensive studies, principally by
the Westinghouse R&D Center, and supported by
others.
The progress of 10 years of effort has resulted in
Union Carbide Corp. building a 1200-MT/year silicon
production plant (now in production) in which the silane
is deposited as silicon in Komatsu chemical vapor
deposition reactors. (These reactors were also
operated in conjunction with the pilot plant.) The silicon
produced is of very high purity. It is expected that the
cost will be less than Siemens-process-produced
silicon, but not as much less as with a FBR.
Silane Process
(Union Carbide Corporation)
Preparation of Silane from Metallurgical-Grade
Silicon and Conversion of Silane into
Semiconductor-Grade Silicon
The process flow diagram is shown in Figure 7.
The dichlorosilane chemical vapor deposition
SiCI4
DISTILLATION
SiH2CI2
SiHCI3
SiH4
FLUIDIZED
BED
REACTOR
SiHCI3
H2
IMPURITIES
SiHCI3
REDISTRI·
BUTION
REACTOR
SiH2CI2
FLUIDIZED
BED
REACTOR
SiHCI3
SiHCI3
SiCI4
SiH2CI2
DISTILLATION
~~XDE L..-
S
_iC...,:14=--
REDISTRI·
BUTION
REACTOR
---I
Si
Figure 7. Silane Process Flow Diagram
17
SILICON
GRANULES
DISTILLATION
Past Accomplishments
• Engineering model FBR successfully operated at
Washougal, Washington:
• Process feasibility demonstrated in laboratoryscale experiments for silane production and for
silane conversion to silicon by thermal
decomposition in free-space reactors and FBRs.
41
e
• Practicality of reactors and process steps for
continuous, steady-state silane production
demonstrated in scaled-up process development
units.
41
• Thermodynamic values and reaction kinetics
measured in hydrochlorination and redistribution
reactors.
Process operating parameters assessed with a
6-in.-diameter FBR.
Long-duration runs with modified reactor
extended range to continuous operation for
66 h using 20% silane.
Conditions for continuous, steady operation
and for producing high-purity silicon remain to
be defined.
• FBR research at JPL:
• Vapor-liquid equilibria calculated for chlorosilanes.
41
• Detailed process design for 100-MT/year EPSDU
completed (silane portion).
• Engineering design and economic analysis for
1000-MT/year plant completed (silane portion).
41
Characterization for high-silane-eoncentration
operation extended to 100 % silane using
6-in.-diameter reactor and cooled distributor
plate (see Figure 8).
Operation at various high silane concentrations
demonstrated, e.g., 80% silane for 3 h at
deposition rate of 3.5 kg/h.
• Hydrochlorination reaction characterized in
extended studies (Solarelectronics, Inc.) consists
of:
41
41
41
Equilibrium constants and reaction kinetics
measured over wide temperature and pressure
ranges.
Reaction rate equation and reaction model
formulated.
Suitable materials of construction for reactor
determined.
• Conventional free-space reactor for conversion of
silane to silicon discarded because of unsuitable
fine powder product.
• FBR R&D plan formulated.
More Recent Accomplishments
• Low-cost silane process technical readiness
demonstrated by operation of pilot plant (EPSDU)
at Washougal, Washington:
41
41
41
High-purity silane produced under optimizedcontinuous steady-state conditions.
High-purity silicon produced using Komatsu
chemical-vapor-deposition reactors.
Operating data obtained for design and
operation of production plant.
Figure 8. JPL Research Fluidized Bed Reactor
18
• Feasibility of periodic removal of product
demonstrated.
• Practicality of redistribution and deposition
reactors as well as integrated unit demonstrated
in scale-up process development unit (see
Figure 10).
• Scavenging mechanism describing high bed
deposition and very low fines formation
demonstrated.
• Redistribution reactor with Dowex catalyst
characterized for kinetics and conversion yield.
• Fine particle growth in silane free-space reactor
at California Institute of Technology:
• Dichlorosilane characteristics determined to
assess hazard:
• Aerosol theory modified to describe formation
and growth of fine particles in silane pyrolysis
system.
• Autoignition, explosion severity, hydrolysis,
and explosive output measured;
dichlorosilane found to be considerably
more hazardous than hydrogen or
trichlorosilane.
• Reactor developed and experimental
conditions investigated with the objective of
growing silicon particles to size suitable for
seed in an FBR.
• Process redesigned, including elimination of
dichlorosilane storage, to minimize
problems.
• Purity of product from both facilities:
• Intermediate and advanced design deposition
reactors studied:
• Cleansing procedures established to ensure
purity of initial seed particles.
• Quartz liners installed in FBR to prevent
contamination from reactor wall.
• Objectives developed from economic goal.
A 2 gh-1 cm-1 deposition rate, 40 mol %
conversion yield, and 60 kWh/kg energy
use.
• Final purity needs to be determined.
• Results from intermediate reactor studies
indicated reactor modifications required:
Dichlorosilane CVD Process
(Hemlock Semiconductor Corporation)
• Deposition rate, conversion yield, and
energy usage objectives achieved
separately.
Preparation of Dichlorosilane from
Metallurgical-Grade Silicon and Deposition of
Semiconductor-Grade Silicon from
Dichlorosilane
• HCI etching step introduced to remove
silicon wall deposits.
• Results from advanced reactor with cooled
wall showed all objectives achievable by
optimizing conditions.
The process flow diagram is shown in Figure 9.
• Process feasibility for conversion of metallurgicalgrade silicon to dichlorosilane and for chemicalvapor deposition of dichlorosilane demonstrated.
• Semiconductor-grade purity demonstrated
for product silicon:
BORON
REMOVAL
Si
SiHCI3
SiH2CI2
BORON IMPURITIES
Si CI4
SiHCI3
SiH2CI2
FLUIDIZED
BED
REACTOR
SiHCI3
SiCI4
SiH2CI2
SiHCI3
SiH2CI2
SiH2CI2
HCI
DISTILLATION H2
REDISTRIBUTION
REACTOR
DISTILLATION
"-MET·
GRADE
Si
SiHCI3
SiCI4
---'SiHCI3
SiHCL3
SiCI4
SiH2CI2
SiCI4
Figure 9. Dichlorosilane CVD Process Flow Diagram
19
Figure 10. Process Development Unit
• Boron, phosphorus, carbon, and metal
concentrations commensurate with, or less
than, those of commercial semiconductorgrade silicon.
determination of impurity effects on material and solarcell properties. The approach was to characterize
materials and solar cells prepared from deliberately
doped ingots.
• Energy conversion efficiency of photovoltaic
cells made from this material shown to be
equivalent to that of baseline cells fabricated
from commercial semiconductor-grade
silicon.
The studies included the following parameters:
• Deep-level impurity concentrations in singly and
multiply doped ingots.
• Methods and growth rates for single crystals .
• EPSDU sized to 100-MT/year outlined.
• Melt-replenishment methods.
• Economic analysis for a 1000-MT/year plant
indicated $34.80/kg (1985 dollars) for product
cost (without profit).
• Base dopant type and concentration (resistivity).
• Anisotropic distributions of deep-level impurities.
Effects of Impurities in Silicon
• Grain boundary structure (impurity interactions in
polycrystalline cells).
The objective was to relate silicon purity to the
economics of silicon refinement processes by the
CD
20
Thermal and gettering treatments during cell
processing.
been incorporated into a slide rule.) The impurity limits
can be relaxed considerably for a near-equilibrium, unidirectional solidification in which liquid-solid impurity
segregation is the concentration controlling factor.
Unfortunately, these studies have not been carried
experimentally into various forms of silicon ribbon and
solar cells made from silicon ribbons.
• High-temperature aging of cells.
•
Effects on high-efficiency PV solar cells (minor
activity).
Tolerable concentrations of impurities were found to
be determined by the effects of: (1) specific metallic
impurities on the performance of cells, (2) boron and
phosphorus on resistivity and single-crystal yield in
ingot growth, and (3) total impurity concentration in the
silicon on single-crystal growth yield. The effects of
impurities on cell efficiency were related empirically
through the base diffusion length and were shown to
be species-dependent. For example, the concentrations
causing a 10% decrease in efficiency ranged from
0.1 ppba for tantalum to 10 ppma for copper (see
Figure 11). The tolerable concentrations in the polycrystalline silicon are dependent on the methods used
for crystal growth. Crystalline structure breakdown
occurs when the total impurity concentration exceeds a
critical value for a particular set of conditions of ingot
or sheet growth. For example, the critical concentration
is in the range of 200 to 500 ppma when a 1O-cmdiameter Cz crystal is pulled at a rate of about 8 cm/h.
(The available database for these trade-off studies has
In a fundamental approach to the impurity effects
problem, a transmission-line-equivalent circuit model
was developed by C.T. Sah to compute the exact
steady-state characteristics of one-dimensional silicon
solar cells. The model was used for calculations of
limiting concentrations of specific impurities for defined
cell structures and efficiency. A model was also
developed to study the effects of defects across the
back-surface-field junction on the performance' of highefficiency cells. A new theory to distinguish an
acceptor-like deep level from a donor-like deep level
was developed using measured values of the thermal
emission and capture cross sections. This theory also
provides information concerning the lattice distortion
around an impurity atom before and after the capture
or emission of an electron or hole at the impurity
center.
METAL IMPURITY CONCENTRATION (ppms)
10-5
10-4
10-3
10-2
10-1
1.0
w
z
iii
en
1.0 ~-----~=---_::::--"""==:---"""==:----~
<t
al
"#:.
>=
~
w
0.8
0.6
Fe
C3
u::
tt
0.4
Cl
w
N - TYPE SILICON
N
~
2
a:
~
0.2
0.0 L-.L--l-L...I.l--L-L-Ul_L-J.-Lll-....L.....L.Lll-----l-----l-Ll.l-...l-...LJ...Ll-----l-----l...LLJ
10 11
1012
1013
1014
1015
1016
1017
1018
METAL IMPURITY CONCENTRATION (atoms/cm 3 )
METAL IMPURITY CONCENTRATION (ppms)
10-5
10-4
10-2
10.3
10-1
1.0
w
z
:::; 1.0
w
en
<t
al
~
>=
u
z
w
0_8
0.6
C3
u::
u..
w
0.4
Cl
W
N
:::;
<t
P - TYPE SILICON
0.2
2
a:
0
Z
0.0
10 11
1012
1013
1014
1015
1016
1017
1018
METAL IMPURITY CONCENTRATION (atoms/cm 3 )
Figure 11. Impurity Concentration Effects on Solar Cell Efficiencies, n- Type and p- Type Silicon
21
• Only about 10% variation in cell performance is
attributable to nonuniform impurity
concentration.
Conclusions
• Decrease of cell conversion efficiency is uniquely
dependent on specific impurities and their
concentrations.
Ell
• General empirical model correlates measured
impurity effects on cell performance.
Research and Development Needs
• Significantly less performance reduction caused
by most deleterious elements in n-base cells.
• Complete understanding of silicon particle growth
mechanisms and particle size control in a
fluidized bed reactor.
• Reduction of electrical activity of impurities by
precipitation in vicinity of grain boundaries in
polycrystalline material is dependent on impurity
diffusion coefficients.
• Establish product purity and other characteristics
of FBR-produced silicon.
Ell
• Removal of impurity elements by thermal and
gettering procedures is dependent on diffusion.
Ell
Ell
Model and data indicate that higher-efficiency
cells are increasingly sensitive to impurities.
Complete characterization of JPL FBR for
operation using high silane concentrations and
producing semiconductor-grade silicon.
• Establish engineering design and operation of
FBR for long-duration steady-state production of
semiconductor-grade silicon.
Composition of acceptable polysilicon feedstock
depends on factors of single-erystal growth
technique, melt-replenishment strategy, and cellprocessing sequence.
• Verify low-cost production of polysilicon in Union
Carbide Corp. silane/FBR EPSDU.
Long-term, ambient-temperature cell
performance is not significantly changed further
by presence of deep-level impurities.
• Define effects of silicon produced by the
silane/FBR process on solar-eell efficiency.
22
Si Iicon Sheet
In 1975, all of the silicon sheet used for
manufacturing semiconductor devices, including
solar cells, consisted of wafers sliced from
Czochralski (Cz) crystals by diamond-embedded
inside-diameter (ID) saws. This semiconductorindustries technology had evolved to a level of
sophistication that was adequate for solar cells for
use on spacecraft. However, the cost of producing
these wafers was prohibitive for large-scale
terrestrial PV uses. The cost of the silicon material
and the processes required to produce the thinsheet constituted the majority of the manufacturing
costs of the early terrestrial PV modules. It was
known that solar cells could be made from silicon
less perfect than single-crystal Cz wafers, but the
effects on solar cell performance were not
understood. Potentially less expensive silicon sheet
had been experimentally grown in a laboratory;
however, the quality of solar cells that might be
made from these new forms of silicon was
unknown. Therefore, an extensive investigation of
silicon materials, its processing into thin sheets, and
its characteristics for PV use was initiated by the
FSA Project.
ED
Develop growth and slicing equipment to meet
these criteria.
ED
Verify equipment performance.
Silicon sheet of quality suitable for fabrication
into solar cells with efficiencies up to 20 %.
ED
Sheet at a cost compatible with FSA goals.
Project Activities: 1975 to 1981
In 1975, under the stimulus created by the need
for large-area low-cost silicon sheet and the
availability of Government funds, several more
growth techniques than mentioned above were
devised. The emphasis was on ribbon technologies
because they seemed to offer the greatest potential
for inexpensive PV power. However, ingot
The key technical issues are to understand the
technologies so that sheet growth can be controlled
to meet:
ED
Material stability with little stress/strain.
Direct growth of silicon ribbons that did not
require cutting to obtain the appropriate thickness for
solar cells obviously was a way to eliminate costly
wafering and silicon-material waste. The growing or
casting of silicon ingots in rectangular shapes so that
they could be cut into square or rectangular wafers,
for more efficient solar cell packing into PV modules,
also seemed to be economically plausible. However,
if cast-ingot technology was to be cost-effective, the
slicing of the large ingots into wafers had to be
accomplished more economically than with
contemporary single-blade ID saws. A common
characteristic of these new silicon sheet forms was
that they were not ideal, with imperfections and
impurities of greater densities and more uneven
distributions than those of Cz crystals. The electronic
properties of these less-than-perfect forms of sheet
were not well known in the 1970s, so many questions
had to be answered. What would be the effect on
solar cell performance? How would performance be
affected by the material purity, the size and
characteristics of the grains, and the quality of the
crystallographic structure? What quality of silicon
sheet could be obtained consistently from the various
sheet-growth techniques? How would the cost to
produce sheet using each specific technique vary
with sheet quality? To answer these questions,
considerable research was needed and
manufacturing cost-analysis capabilities had to be
developed that could be used to periodically assess
the progress of the competing technologies.
The goals now are to establish crystalline silicon
growth technology appropriate for:
ED
ED
Based upon the silicon sheet R&D performed
during the early 1960s, considerable potential existed
in 1975 for lower-cost silicon sheet that was
surmised to be of adequate quality for terrestrial
solar cells, i.e., growth of ribbons, casting of sili.con,
and advanced (less expensive) Cz and float-zone
(FZ) crystal-growth techniques. Three new ribbon
growth techniques, by which silicon is grown directly
to the thickness of a solar cell, had been conceived
and their feasibility had been demonstrated
experimentally: the dendritic-web, Stepanov, and
edge-defined film-fed growth (EFG) methods.
The goals were to:
Develop silicon sheet growth technologies to
meet specific cost and quality criteria (including
ribbon growth, ingot growth, and ingot slicing).
A sheet quality with acceptable crystallographic
and impurity characteristics, i.e., type, quantity,
and uniformity of distribution.
Background
The objective of the silicon sheet activity is to
develop a sheet growth technology that will meet the
cost and quality requirements for large-scale PV
applications. Initially, the objective had been to
define, develop, and demonstrate process
technologies capable of producing large-area silicon
sheet material for production of low-cost solar arrays
at a price of $0.50 per peak watt (1975 dollars) by
1986.
ED
ED
High-area throughput rates that meet quality
and economic needs.
23
rates'are'faster. However, they are polycrystalline.
The quality, size, and uniformity of grains in the ingot,
plus the non-uniform distribution of impurities, are
parameters that influence the electronic quality of
specific cast ingots. The efficiencies of solar cells
made from these materials are generally less than
those of Cz cells, and are highly dependent upon
the ingot quality. Goals and status are shown in
Tables 6 and 7.
technologies were also funded because there was
greater assurance that these efforts could succeed,
An initial plan was devised that included a four-phase
effort: research and development on sheet-growth
methods (1975-1977); advanced development of
selected growth methods (1977-1980); prototype
production development (1981-1982); development,
fabrication, and operation of growth production plants
(1983-1986),
Concurrently, with the early phases, a cost-tomanufacture economic analysis capability was
developed and refined to maturity. Periodic
assessments of the economic value of each sheet
technology were performed, based upon the latest
progress of sheet and other PV technologies. The
sheet technologies showing the greatest promise for
economic viability were selected for advanced
development, and the less promising were dropped.
The silicon sheet technologies funded by FSA during
the late 1970s are shown in Tables 3 and 4. The
lists of active contracts in June 1977 and September
1978 show that in this period, the less promising
ribbon techniques were dropped, and that new
advanced Cz growth efforts (and die and container
material studies) were initiated. Early in 1981, a
major review of the remaining sheet techniques was
performed. The dendritic web, EFG, and silicon-onceramic (SOC)' efforts continued to be supported.
Also, the advanced Cz (by Kayex Corp.), heatexchange method (HEM), a Semix, Inc., casting
process, and multiwire ingot slicing received
continued support.
Three ingot slicing techniques on which FSA has
sponsored R&D are ID, multiblade, and multiwire
(shown in Figures 15, 16 and 17). The ID and multiwire saws have cutting materials embedded in a
softer matrix and are subject to wearout as the cutting
particles are dulled or eroded away. The multiblade
saw and a multiwire saw (also funded by DOE) use a
slurry containing abrasive particles. In all cases, cooling of the cutting area with a fluid is required. The
sawing occurs by the tearing away of silicon, causing
surface damage to the sawed wafer. The surface
damage is different for each sawing technique; in
each case it must be removed, usually by etching the
wafer in acid. Goals and status are shown in Table 8.
The most advanced and still the most promising
ribbon technologies are the EFG methods and the
dendritic web methods (shown in Figures 18 and 19),
In the EFG method, the molten silicon moves
upward between parallel walls of a graphite die by
capillary action. A silicon seed is lowered to make
contact with the molten silicon in the top of the die,
and is then pulled upward. The shape of the resulting
ribbon grown on the seed is defined by the shape of
the top of the die. Relatively large quantities of ribbon
can be grown quickly, especially by pulling several
ribbons simultaneously from the same melt. Mobile
Solar Energy Corp. devised, with its own funds, and
is producing nine ribbons from one melt by its
"nonagon" concept (a tubular shape with nine flat
sides). The quality of the EFG ribbon is its major
limitation: it produces solar cells with lower efficiencies than are possible with Cz or FZwafers. The
molten silicon reacts with the graphite dies and, as
the silicon cools, silicon carbide precipitates are incorporated into the ribbon. The result is a ribbon that has
a disrupted crystalline structure with uneven impurity
distributions. Goals and status are shown in Table 9.
The Cz growth technique is shown in
Figure 12. Semiconductor silicon chunks, the
"charge," are melted in a quartz crucible within a
sealed grower furnace filled with argon. A silicon
"seed" of the appropriate crystallographic orientation
is lowered until it is in contact with the molten silicon
surface, and is then withdrawn slowly. The singlecrystal silicon that grows on the seed has the same
crystallographic structure as the seed, as long as the
furnace conditions and seed withdrawal are properly
controlled. The dimensions of the Cz crystal or ingot
are determined by the quantity of the silicon charge
and the rate of seed withdrawal. The dopant and
impurity levels in the crystal are determined by the
trace elements in the charge, except that additional
oxygen and carbon are introduced during the growth
process from the quartz (Si0 2) crucible and the
graphite furnace components. Control of impurities,
structural defects, and their uniformities have been
improved considerably by careful analysis,
experimentation, and improved grower design and
operational control. Goals and status are shown in
Table 5.
Ribbon grown by the dendritic,web method is of
higher quality than EFG ribbon and other ribbons. Its
electronic quality can be almost as good as that of
Cz wafers, except that it has a "twinning plane" in
the plane of the ribbon throughout the ribbon length.
By precise control of temperatures and temperature
gradients, during and after seeding and growth
initiation, parallel silicon filaments, called
"dendrites," can be grown in the silicon melt. Under
carefully controlled withdrawal, the dendrites
continue to grow and a film of molten silicon forms
and solidifies between them. The only solid material
that the molten silicon contacts is the quartz crucible
which minimizes contamination and results in a
Cast-silicon ingots (Figures 13 and 14) have the
advantage of being cast in such a form that square or
rectangular wafers can be sliced directly from the
ingot. Some cast ingots are large enough that they
can be slabbed into four or nine square ingots before
being sliced into wafers. Cast ingots are less
exoensive than Cz ingots because their solidification
24
Table 3. Silicon Sheet Contractors: June 19 77
Technology Area
Contractor
Ribbon Growth Processes
Mobil-Tyco Solar Energy
Waltham, Massachusetts
(JPL Contract 954335)
Edge-defined, film-fed growth
IBM
Hopewell Junction, New York
(JPL Contract 954144)
Capillary action shaping technique
RCA
Princeton, New Jersey
(JPL Contract 954465)
Inverted Stepanov growth
Univ. of So. Carolina,
Columbia, South Carolina
(JPL Contract 954344)
Web-dendritic growth
Motorola
Phoenix, Arizona
(JPL Contract 954376)
Laser zone ribbon growth
Westinghouse Research
Pittsburgh, Pennsilvania
(JPL Contract 954654)
Dendritic web process
Sheet Growth Processes
Honeywell
Bloomington, Minnesota
(JPL Contract 954356)
Dip coating of low-cost substrates
Rockwell
Anaheim, California
(JPL Contract 954372)
Chemical vapor deposition on low-cost substrates
General Electric
Schenectady, New York
(JPL Contract 954350)
Chemical vapor deposition on floating silicon
substrate
Univ. of Pennsylvania,
Philadelphia, Pennsylvania
(JPL Contract 954506)
Hot-forming of silicon sheet
Ingot Growth Process
Heat-exchanger ingot casting a
Crystal Systems
Salem, Massachusetts
(JPL Contract 954373)
Ingot Cutting
Crystal Systems
Salem, Massachusetts
(JPL Contract 954373)
Multiple wire sawing a
Varian
Lexington, Massachusetts
(JPL Contract 954374)
Breadknife sawing
aSingle contract provides for both ingot casting and mUltiple wire sawing.
25
Table 4. Silicon Sheet Contractors: September 1978
Contractor
Technology Area
Shaped Ribbon Technology
Mobil-Tyco Solar Energy
Waltham, Massachusetts
(JPL Contract 954355)
Edge-defined film-fed growth
Motorola, Inc.
Phoenix, Arizona
(JPL Contract 954376)
Ribbon growth, laser zone regrowth
Westinghuuse Research
Pittsburgh, Pennsylvania
(JPL Contract 954654)
Dendritic web process
Supported Film Technology
Honeywell Corp.
Bloomington, Minnesota
(JPL Contract 954356)
Silicon on ceramic substrates
RCA Labs
Princeton, New Jersey
(JPL Contract 954817)
Epitaxial film growth on low-eost silicon substrates
Ingot Technology
Crystal Systems, Inc.
Salem, Massachusetts
(JPL Contract 954373)
Heat exchanger method (HEM), cast ingot, and
multiwire fixed abrasive slicing
Kayex Corp.
Rochester, New York
(JPL Contract 954888)
Advanced Cz growth
Siltec Corp.
Menlo Park, California
(JPL Contract 954886)
Advanced Cz growth
Texas Instruments
Dallas, Texas
(JPL Contract 954887)
Advanced Cz growth
Varian Vacuum Division
Lexington, Massachusetts
(JPL Contract 954374)
Multiblade slurry sawing
Varian Vacuum Division
Lexington, Massachusetts
(JPL Contract 954884)
Advanced Cz growth
Die and Container Materials Studies
Battelle Labs
Columbus, Ohio
(JPL Contract 954876)
Silicon nitride for dies
Coors Porcelain
Golden, Colorado
(JPL Contract 954878)
Mullite for container and substrates
Eagle Picher
Miami, Oklahoma
(JPL Contract 954877)
CVD silicon nitride and carbide
RCA Labs
Princeton, New Jersey
(JPL Contract 954901)
CVD silicon nitride
Tylan
Torrance, California
(JPL Contract 954896)
Vitreous carbon
26
ADVANCED CZOCHRALSKI
CRYSTAL GROWTH SYSTEM
CZOCHRALSKI INGOT
INGOT PULL AND
ROTATION
-----<'"1
MECHANISM
MULTIPLE INGOT CAPABILITY
DOOR
CRYSTAL REMOVED
HOPPER WITH SILICON INSERTED
LOWERED TO CRUCIBLE - UNLOADED
SECOND CRYSTAL GROWN
CYCLE REPEATED
SEED
CRYSTAL - - - - - h i ' k i l 1
FURNACE
CHAMBER
ISOLATION
VALVE
--~~
MOLTEN
SILICON
GROWING
CRYSTAL
Figure 12. Advanced Czochra/ski Crystal Growth by Hamco Division (Kayex Corp.)
HEAT EXCHANGER METHOD
VIEW PORT
INSULATION
1:--, 'L,u"'l.Jj:--HEATING ELEMENT
POWER LEADS~;I.';..-;e'v.SILICON MELT
CRUCIBLE
• TO VACUUM PUMP
PYROMETER
SEED CRYSTAL
CHAMBER
HELIUM GAS
Figure 13. Ingot Casting by Heat Exchanger Method by Crystal Systems, Inc.
Figure 14. Ingot Casting by Semicrystalline Casting Process by Semix, Inc.
27
Table 5. Advanced Czochralski Crystal Growth by Hamco
GOALS:
STATUS
R&D NEEDS
" 150 kg OF INGOTS/
CRUCIBLE
" 150 kg OF INGOTS/
CRUCIBLE
"INGOT QUALITY:
STRUCTURAL PERFECTION
" 15-cm-DIA INGOT
" 15-cm-DIA INGOT
" GROWTH PARAMETER
OPTIMIZATION
" 2.5 kg/hr THROUGHPUT
" 2.2 kg/hr THROUGHPUT
1986 INGOT ADD-ON PRICE'
(1980 $/W p)
" AUTOMATION
" UNDER DEVELOPMENT
AT END OF CONTRACT
(LATER ACHIEVED BY HAMCO)
" 15% ENCAPSULATED SOLAR
CELL EFFICIENCY
" 15% SOLAR CELL
EFFICIENCY' ,
"FSA PRICE
ALLOCATION 0.112
"90% YIELD
"86% YIELD
" SAMICS
PROJECTION 0.162
'ASSUMES 0.73 m2 OF WAFERS/kg OF INGOT
"ENCAPSULATED CELL EFFICIENCY N.A.
Table 6. Ingot Casting by Heat Exchanger Method by Crystal Systems, Inc.
GOALS:
STATUS:
R&D NEEDS
" 30 em x 30 em x 15 em
SHAPED INGOT (35 kg)
" 34 em x 34 em x 17 em
" IMPROVED INGOT QUALITY
" >90% SINGLE CRYSTAL
" > 95% SINGLE CRYSTAL
" 2.0 kg/hr GROWTH RATE
" 1.9 kg/hr GROWTH RATE
" INCREASED INGOT YIELD
1986 INGOT ADD-ON PRICE"
(1980 $/Wpl
" 15% ENC. SOLAR CELL
EFFICIENCY
" 15% SOLAR CELL
EFFICIENCY
" FSA PRICE
ALLOCATION 0.194
">95% YIELD
">95% YIELD
" SAMICS
PROJECTION 0.17
SHAPED INGOT (35 kg)
" ASSUMES 1 m 2 OF WAFERS/kg OF INGOT
Table 7. Semicrystalline Casting Process by Semix
STATUS:
R&D NEEDS
" 30 em x 30 em x 19 em
SHAPED INGOT (40 kg)
" 20 x 20 x 15
(14 kg)
" LARGE QUANTITY OF
LARGE AREA CELLS AT
15%
" SEM ICRYSTALLINE
" SEMICRYSTALLINE
" DEMONSTRATION OF
HIGH THROUGHPUT
PROCESSING
" 1 5% SOLAR CELL
EFFICIENCY
"
" 98% INGOT YIELD
" 83% INGOT YIELD
"FSA PRICE
ALLOCATIONS 0.256
" 2.3 kg PER HOUR
CRYSTALLIZATION
RATES
" 2.3 kg PER HOUR
" SEMIX PROJECTION
0.115
GOALS:
15% SOLAR CELL
EFFICIENCY
"USING 0.81 m 2 OF WAFERS/kg OF INGOTS
28
1986 ADD-ON PRICE"
(1980$!Wp)
SILICON TECHNOLOGY CORP
1.0.* SAW
IMBEDDED
DIAMOND
PARTICLES
'INNER DIAMETER
Figure 15. Inside Diameter Ingot Saw by Silicon Technology Corp.
HOFFMAN
MULTIBLAOE SAW
SLURRY WITH
CUTIING
L - _ - \ -.....
COMPOUND
Figure 16. Multiblade Saw by P. R. Hoffman Co.
CRYSTAL SYSTEMS, INC.
MULTI-WIRE SAW
Or)
/
ROCKING
INGOT
HOLDER
• Figure 17. Multiwire Saw by Crystal Systems, Inc.
29
Table 8. Advanced Wafering Technologies by Crystal Systems Inc., P.R. Hoffman, and
Silicon Technology Corp.
GOALS
STATUS
(BEST SINGLE DEMONSTRATION)
15 em DIA
10 x 10 em square
15 em DIA
10 x 10 em
20 WAFERS/em
OF INGOT
25 WAFERS/em OF
INGOT
17 WAFERS/em
25 WAFERS/em
0.5 WAFER/min
1 WAFER/min
0.42 WAFER/min
1.25 WAFER/min
0.82 m 2/kg'
1.0 m 2 /kg'
0.69 m 2/kg"
1.0 m2/kg
95% YIELD
95% YIELD
95%
95%
ADDITIONAL R&D
REQUIRED
• DEVELOP FUNDAMENTAL UNDERSTANDING
OF WAFERING PROCESS, FOR
INCREASING THROUGHPUT
• INCREASED WAFER THROUGHPUT
• REDUCED EXPENDABLE COSTS
• SENSITIVITY TO EXTENDED OPERATION
1986 WAFER ADD-ON PRICE
(1980 $/W p)
• FSA PRICE ALLOCATION
0.081 (15 em DIA)
• SAMICS 0.110 (15 em DIA)
'NO LONGER UNDER DEVELOPMENT
"SQUARE METER OF WAFERS/KILOGRAM OF INGOT
Project Activities: 1981 to Present
clean, smooth ribbon. The dendrites are easily
removed after the ribbon has cooled. The major
disadvantage of this method is the low production
rate of quality ribbon. Significant difficulties have
been encountered as faster growth rates and
sustained operation for many hours have been
attempted. The problems of growing low-stress
ribbons at economical growth rates continue to
plague the web efforts. Goals and status are shown
in Table 10.
With a change in Project directions, R&D on
specific growth technologies was then replaced by
generic growth R&D, More recently, a concerted
effort to increase the area growth rates ot hlghquality dendritic web are now supported by a
consortium of utilities, DOE and Westinghouse.
The status of web and EFG in 1985 is shown in
Table 13.
As time passed, ribbon-buckling problems and
reduced ribbon quality that were encountered when
higher ribbon growth rates were attempted led to the
conclusion that a more fundamental understanding of
the impediments to high-speed silicon ribbon growth
and quality was required. This resulted in the
establishment of an interactive university-industryGovernment effort that continues to actively pursue
theoretical and experimental ways to achieve highspeed, high-quality silicon crystal growth. Computer
models have been developed and are being used to
study the stresslstrain relationship in the silicon
solidification zone and the nearby areas in the
ribbon. The analytical results are then experimentally
evaluated on ribbon-growth equipment at
Westinghouse and Mobil Solar. In support, hightemperature silicon material properties are being
determined and sophisticated instrumentation and
measurement techniques continue to be developed.
Progress and problems of the various efforts are
discussed at frequent meetings and the results
iterated into the appropriate activities. A low-angle
(horizontal) silicon sheet growth effort was also
initiated to explore the potential value of this very
high-speed-ribbon-growth technique (Figure 21 and
Table 14). This technique has the potential to evolve
into a practical growth technology, but is not now
funded by DOE because of a shortage of funds. It
was funded in an attt3mpt to gain generic information
about high-speed silicon growth.
The silicon-on-ceramic (SOC) technique was the
third promising non-ingot growth method. It was
unique in that it had high growth rates; but the
efficiency of solar cells made from it was low.
Consequently, when the need for higher efficiency
became widely accepted and the Project funds were
cut, SOC efforts were dropped (see Figure 20 and
Table 11).
In 1981, the direction of the National Photovoltaics
Program was altered to emphasize long-term, highrisk research efforts, and funding was significantly
reduced. Concurrently, there was a growing
awareness that PV module efficiencies had to be
raised considerably if PV was to be cost effective for
large utility applications. Higher-efficiency modules
and solar cells require higher-quality silicon sheet.
Consequently, the production-development-oriented
third an,d fourth phases were discontinued, and ingot
growth, ingot slicing, and ribbon technologies other
than dendritic web and EFG were no longer funded.
The status of the sheet technologies that were
dropped because of programmatic changes are
shown in Table 12, as they were at time of contract
completion. An exception, low-angle silicon sheet
(LASS) growth method, was funded for about 2 years.
30
3-to 1O-em WIDE RIBBON GROWERS
f
EFG* RIBBON
DIE
CAPILLARY
ACTION
MOLTEN
SILICON
'EDGE-OEFINED FILM-FED GROWTH
Figure 18. Edge-Defined Film-Fed Ribbon Growth by Mobil Solar Energy Corp.
WEB DENDRITIC
SHEET GROWTH
Figure 19. Dendritic-Web Ribbon Growth by Westinghouse
t
:~t
__ _
MOVABLE DISPLACER
SILICON
MELT
GRAPHITE ROLLERS
Figure 20. Silicon-on-Ceramic by Honeywell
31
Table 9. Edge-Defined Film-Fed Ribbon Growth by Mobil Solar Energy Corp.
GOALS:
STATUS:
R&D NEEDS:
• SIMULTANEOUS GROWTH OF'
4 RIBBONS, 10 em WIDE EACH,
AT 4 em/min FOR 8 hr
WITH AUTOMATIC CONTROLS
• 40 em 2/min (10 em WIDE'
x 4.0 em/min) (4 hr DEMO
OF ONE RIBBON)
• IMPROVED SHEET QUALITY
• MULTIPLE RIBBON GROWTH*
STATION
• DEMONSTRATED'
(5 RIBBONS AT 5 em WIDE)
(3 RIBBONS AT 10 em WIDE)
1986 SHEET ADD-ON PRICE
(1980 $/W p )
• 0.20 mm RIBBON THICKNESS
Ell
0.15-mm RIBBON THICKNESS
• 12% ENCAPSULATED SOLAR
CELL EFFICIENCY
Ell
>12% SOLAR CELL"
EFFICIENCY
Ell
INCREASED THROUGHPUT
GROWTH SPEED
• FSA PRICE
ALLOCATION 0.205
• SAMICS'
PROJECTION 0.194
• 80% YIELD
'MOBIL DEVELOPING COMPETITIVE NONAGON EFG TECHNOLOGY
"ENCAPSULATED CELL EFFICIENCY N.A.
Table 10. Dendritic-Web Ribbon Growth by Westinghouse
GOALS (LATE 1986):
R&D NEEDS:
e 30-cm2/min AREA
STATUS:
e 13-cm2/min AREA GROWTH RATE
GROWTH RATE FOR 50 METER·LONG
FOR SHORT RIBBON LENGTHS AND
CRYSTAL AND CONSTANT MELT LEVEL
GROWTH RATE
e INCREASED THROUGHPUT
e 17.5% SOLAR CELL EFFICIENCY
e PROCESS AUTOMATION
e 16.9% CELL EFFICIENCY
e AUTOMATION BEING INCORPORATED
e 8-hr DEMONSTRATION RIBBON GROWTH
FOR 8-3/4 hr AT CONSTANT
MELT LEVEL
e GROWTH AND STABILITY
1986 SHEET ADD-ON PRICE
(1980 $/W p)
e FSA PRICE
ALLOCATION 0.292
e SAMICS
PROJECTION 0.160
Table 11. Silicon-on-Ceramic by Honeywell
GOALS:
•
<$6/m 2
•
350-em 2/min
SUBSTRATE COST
AREA GROWTH
RATE (2 SHEETS)
• 0.10-mm FILM THICKNESS
• 11% ENCAPSULATED SOLAR
CELL EFFICIENCY
• 95% YIELD
STATUS:
ADDITIONAL R&D REQUIRED:
• MULLITE SUBSTRATE IDENTIFIED
<$6/m 2 PROJECTED
• IMPROVED SHEET QUALITY
e INCREASED THROUGHPUT
• 60-em 2/min AREA
GROWTH RATE (1 SHEET)
• OPTIMIZED CELL DESIGN
• <0.10-mm FILM THICKNESS
1986 SHEET ADD-ON PRICE
(1980 $/W p)
• 10.5% SOLAR CELL"
EFFICIENCY ON DIP-COATED SOC
e FSA PRICE
ALLOCATION 0.190
e SAMICS
PROJECTION N.A.
'NO LONGER UNDER DEVELOPMENT
"ENCAPSULATED CELL EFFICIENCY N.A.
32
Table 12. Technology Status at Completion of Contract
MATERIAL
SHEET
TECHNOLOGY
THROUGHPUT
(kglhr)tor (m 2Ihr)
DIMENSIONS
(em)
INGOT
TR GOAL
Adv. CZ
UTllI~ATION
m Ikg
SOLAR CELL EFFICIENCY
%AMl
ENCAPSULATED
BARE CELL
ADD-ON
PRICE
SIW p(1980S)
ESTIMATED'
STATUS
TR
ALLOCATION
STATUS
TR GOAL
STATUS
TR GOAL
STATUS
TR GOAL
15 DIA
15 DIA
2.5t
2.2t
n,a.
n.a.
15
13.5
0.112
HEM
30x30x15
34x34x17
1.0t
1.0t
n.a.
n.a.
15
11.5
0.194
Semix Casting
Process
30x30x19
20x20x15
2.3....
2.3....
n.a.
n.a.
15
11.0
0.245'"
Adv. Wafering"
15 DIA &
10xl0 SO
15 DIA &
10xl0 SO
0.53
0.60
0.42
0.75
0.73
1
0.65
1
n.8.
n.a.
n.a.
n.a.
0.081
0.062
3.6
12
4.3
11
9.5
0.190
SHAPED SHEET
LASS
1 @ 15
SOC
2 @ 12.5
x 0.010
3.6
1 @ 10
x 0.010
2.1
4.3
0.3
'HIGHER EFFICIENCY INDIVIDUAL CELLS HAVE BEEN FABRICATED
"DATA GIVEN FOR ID WAFERING. OTHER WAFERING APPROACHES WILL ALTER THE FIGURES
'''ADDED VALUE FOR CASTING AND WAFERING
....CRySTALLIZATION RATE kglhr
Table 13. Technology Status of Active Contracts
SHEET
TECHNOLOGY
MATERIAL
C~RRENT
SOLAR CELL EFFICIENCY
%AMl
ENCAPSULATED
BARE CELL
ADD-ON
PRICE
S!W p (1980S)
TR GOAL
CURRENT
STATUS
TR GOAL
ESTIMATED'
STATUS
TR
ALLOCATION
THROUGHPUT
(m 2/hr)
DIMENSIONS
(em)
UTllI~ATION
m /kg
TR GOAL
STATUS
TR GOAL
CURRENT
STATUS
EFG"
4@ 10WIDE
x 0.020 THICK
3@ 10WIDE
X 0.020 THICK
0.96
0.60
2.15
2.15
12
14
0.205
WEB
-
-
0.18
0.05
<;;;3.2
<;;;3.2
15
16.9
0.292
SHAPED SHEET
'HIGHER EFFICIENCY INDIVIDUAL CELLS HAVE BEEN FABRICATED
"NONAGON RIBBON CONFIGURATION FUNDED BY MOBIL SOLAR HAS HIGHER VALUES.
Table 14. LOW-Angle Silicon Sheet Growth by
Energy Materials Corp.
STATUS (BEST INDIVIDUAL ACHIEVEMENT)
RIBBON
SCRAPER
MELT
RIBBON WIDTH:
15 em
RIBBON THICKNESS:
25 mils
PULL SPEED:
1 RIBBON AT 85 em/min
THROUGHPUT:
450 em 2/min
EFFICIENCY:
12.9%
HEAT FLOW MODIFIER
CRUCIBLE
demonstrated individually. But until a few months
ago, progress had been very slow in the efforts to
achieve all growth requirements simultaneously and
to grow web continuously for hours while meeting the
requirements. To obtain a fresh look at this growth
technique, a special JPL Web Team was formed and
has recently made interesting advancements
(Table 15). During the last few months, Westinghouse
has also made good progress (Table 16). The major
question still remains: How economically can highquality web be grown under production conditions?
Figure 21. Low-Angle Silicon Sheet Ribbon Growth
by Energy Materials Corp.
Of all of the sheet technologies, dendritic web is
viewed today as having the greatest potential for
high-quality sheet with the best economical value.
However, close temperature control of the silicon
liquid-solid interface regions in the grower are
required. By 1981, a number of web-growth
requirements to meet Project goals had been
33
Table 15. JPL Web Team
Objective
The objective of the JPL Web Team is to conduct an independent R&D activity that will increase the
likelihood that dendritic web technology will achieve the DOE goals.
Approach
Team effort consists of a combination of analytical and experimental work.
Operate a modified Westinghouse web growth system.
Measure high temperature physical properties of silicon.
Use improved stress web model to determine a web temperature profile resulting in higher area
growth rates of web with satisfactory quality.
Use thermal model of web to determine a growth system thermal configuration that will yield the
desired web temperature profile.
Use thermal analysis of web-growth interface and the susceptor-crucible melt to determine a
growth system thermal configuration that will improve stability of growth process.
Design, fabricate, and test the new thermal configurations and feed results back into the models.
Accomplishments
High temperature properties of silicon have been measured.
A stress model of growing web has been developed and used.
A thermal model of growth system configuration to control web temperature has been developed.
A thermal model of susceptor, crucible, and melt has been developed.
Westinghouse growth system has been installed with an improved data acquisition system.
Numerous web ribbons have been grown.
Table 16. Dendritic- Web Growth Progress by Westinghouse (Westinghouse data)
1981
1982
7
4
13
8
1977
1978
1979
1980
Area Growth Rate, cm 2/min
1) Transient (Lengths Of Several Centimeters)
2) Ouasi-Steady-State (Lengths Of 30 To 100 cm)
3) Steady-State (Meters Of Length, Hours Of Growth)
2.3
8
23
27
Maximum Undeformed Width, Centimeters
2.4
3.5
4.0
4.4
5.5
2.0
2.7
3.2
4.9
Development
W150-Undeformed Width At 150 IJ.m Thickness
(An Inverse Measure Of Buckling Stress)
1983
Mid
1985
42
6.7
5.8
Maximum Area Throughput For Single Furnace in
5-Day Week (cm 2/week)
Maximum Demonstrated Solar Cell Efficiency,
AM1
1984
9,000 27,000
13
34
14
15
15.5
15.9
16.2
16.9
All three ingot technologies (shown in Figures 12,
13, and 14) are in commercial use today for
terrestrial solar cells. The majority of the wafers for
terrestrial solar cells are still sliced from Cz singlecrystal ingots. In 1975, the solar cells fabricated
specifically for terrestrial use were made from the
circular wafers sliced from Cz cylindrical crystals.
Today, Cz crystals generally are slabbed lengthwise,
converting cylinders into polygons that are then sliced
into polygon wafers, thereby enabling the solar cells
to be more efficiently "packed" into modules. Now,
much larger Cz crystals are grown and multiple
crystals can be grown from the same crucible by
using melt-replenishment. Formerly, a new crucible
was required for each crystal grown.
Progress Since 1975
Ingot Technology
Ell
Cz Ingots
'" Ingot diameters increased from 3 to 6 in.
'" Throughput rates were increased threefold.
'" Ingot quality was improved.
'" Sheet costs were decreased to one-seventh.
Ell
Today, ID sawing remains the predominant
commercial, and most economical, slicing
technology. More economical sawing remains the
major impediment to continued long-term use of ingot
technologies for PV. Studies of the basic
mechanisms involved in the removal of silicon,
including the influence on cutting rates and surface
damage of various fluids, are promising. These
results, as yet, have not been applied in an attempt
to develop a new saw or to improve existing saws.
Cast Ingots
'" Two concepts have proven feasible: HEM
and Semix (semicrystalline).
'" The same two techniques have advanced to
commercial products.
Ell
Wafering
.. Developed and demonstrated increased ingot
slicing capabilities.
The best-quality silicon sheet today is obtained by
FZ crystal growth; the next best is by Cz crystal
growth. Very high-efficiency solar cells have been
made in the laboratory from FZ wafers. However, FZ
technology is not used commercially for PV because
of its high cost. FZ technology development has
never been part of the FSA Project. Excellent
progress in reducing the cost of growing Cz crystals
was made during the late 1970s, primarily by FSAsponsored efforts. During the early 1980s, continuing
progress in improved Cz crystal quality has been
made for and by the integrated'circuit industry to
meet its needs for more compact and sophisticated
devices. The achievement of even higher-quality Cz
crystals by MCz (Cz crystals grown in a magnetic
field), improved understanding of growth phenomena,
and better equipment is continuing for more stringent
future integrated-circuit industry needs. These
activities fit in well with the PV need for higherefficiency cells. However, the major obstacle to
future use of Cz or FZ wafers for PV is the cost of
wafering. The added cost of the wafering step is
incidental to the cost of manufacturing integratedcircuit devices, but it looms as the single most costly
item if Cz or FZ technology is to be used for future
solar cells (one cell/wafer) and modules. A very low
key, but promising effort on understanding the basic
mechanism of silicon fracturing during sawing has
been continuing. With the uncertainties in the
economical value of ribbon growth, it would be
prudent to increase efforts leading to more
economical slicing and an evaluation of the technical
and economical potential of advanced Cz and
advanced FZ crystal growth for higher-efficiency,
low-cost PV cells and modules.
'" Reduced production costs.
Ribbon Technology
Ell
Thirteen ribbon and non-ingot technologies were
investigated.
Ell
Two ribbon technologies continue to be
supported:
'" Dendritic web, by Westinghouse Electric
Corp.
'" EFG, by Mobil Solar Corp. (now funded only
as generic ribbon growth).
'" One ribbon technology (EFG) is now
commercially available. One ribbon
technology (web) is scheduled to be
demonstrated by late 1986.
'" Significant improvements in understanding
high-speed crystalline silicon growth
phenomena are being obtained.
o
Low-angle silicon sheet, by Energy Materials
Corp. (contract ended June 1985).
Recent Achievements
Ell
35
Thin, flat, essentially stress-free dendritic-web
ribbon width has been increased to 6.7 cm
(Westinghouse).
• Using a new grower with dynamic controls,
dendritic-web linear growth rates have been
increa~ed to 3.0 em/min for 150-fLm-thick ribbons
under transient growth conditions (Westinghouse).
• Dendritic-web transient area growth rate for short
ribbon lengths was increased from 27 to 42
cm 2 /min (Westinghouse: see web growth
progress, Table 16).
• Continuous web ribbon growth with melt
replenishment was increased to 7-m length at
7 cm 2 /min using technology developed for FSA
(Westinghouse).
e
Fracture mechanics and surface chemistry
effects on:
G
Device efficiency.
G
Cell and module performance lifetime.
Fabrication and processing rates and
yields.
• Correlation of observed buckling phenomena
and stress model.
• Data on basic silicon material properties (i.e.,
creep behavior and stress moduli) at
temperature range of interest (800 to 1400 0c).
• Correlation of structure (i.e., dislocations and
grain boundaries) effects with generated stress.
• Impurity effects on stress generation.
• A comprehensive plan for increasing ribbon
quality and growth capability has been devised
and is being partially implemented.
• Verification of stress model with experimental
growth systems in industry.
Research and Development Needs
It
Silicon sheet material R&D is needed for ribbon or
ingot in the cost-to-produce range of $90/m 2 , suitable
for production-type processing into high-performance
17 to 20% efficiency solar cells.
Construction of or modification of a growth
system that integrates all information generated
to produce optimum silicon ribbons at high
speed.
Ingot Technology
• Requirements
A 60- to 100-cm/s bulk recombination
velocity.
Effects of impurities on stress/strain in silicon
growth.
• A general thermal stress model of ribbon
growth integrating creep behavior, growth
speed, and ribbon width should be completed
and applied to all ribbon growth processes.
• A research forum and other meetings on highspeed growth and characterization of crystals for
solar cells have resulted in a definition of the
areas requiring research.
e
e
Ribbon Technology
• Microhardness tests of silicon in various fluid
environments has revealed that the surface
hardness of silicon varies apparently as a function
of the fluid's dielectric constant. Tests were
conducted with toluene, acetone, ethanol,
methanol, glycerol, and deionized water
(University of Illinois at Chicago).
Diffusion length equal to or greater than
300 fLm.
Stress/strain effects on ribbon and bulk
silicon during growth.
G
• Integration of the theoretical studies of heat flow
and interface stability by Massachusetts Institute
of Technology (M IT), and plastic flow near the
interface region by Harvard University, were integrated with residual stress data by Mobil Solar.
e
e
It
Improved understanding of the effects of
impurities on the growth of uniformly highquality single crystal ingots.
• Improved understanding of defect formation
during single-crystal ingot growth.
It
Verification of the feasibility of growth
techniques that improve quality of ingots.
It
Improved understanding of slicing phenomena
that will lead to higher quality, more economical
wafers.
• Research Areas
Basics of physics and chemistry of silicon
materials at room temperature and near the
melting point:
36
Current Research and Plans
€I
• Stress and strain relationships in wide, thin
ribbons grown at high speeds:
€I
€I
€I
€I
€I
Develop methods of measuring thermal fields
in growing ribbons.
Develop methods of measuring residual stress
in ribbons.
€I
Develop methods of observing ribbon
deformation in real time.
€I
€I
Model heat flow and interface stability criteria
for ribbon growth.
Test and verify model.
• Sheet quality effects on solar cell performance:
Measure the high-temperature mechanical
properties of silicon.
Develop models for plastic deformation and
elastic buckling in growing ribbons.
€I
€I
Design, fabricate, and test low-stress ribbon
growth systems.
• Impurity effects and interface stability during
ribbon growth:
€I
€I
Modify present ribbon growth processes to use
the data obtained on the above.
Verify models by relating the observed thermal
fields to observed ribbon stress and strain.
€I
Use verified models to define ribbon thermal
fields and growth dynamics that will yield lowstress growth environments.
€I
37
Characterize ribbon material grown under
various conditions.
Fabricate baseline cells on ribbons grown
under various growth-process conditions.
Relate cell performance to ribbon growth
process variables and specific defects.
Optimize growth processes that compensate
for specific ribbon defects.
progress in cell designs and process techniques
were largely empirical, which has been strengthened
by better understanding of the underlying principles
of operation. These activities have led to a general
agreement regarding the factors that limit cell
performance and to those that can be changed to
further increase the performance of crystalline silicon
solar cells.
High-Efficiency
Solar Cells
During the past 10 years, as FSA efforts
continued, it has been perceived that the original
Project goal of 10% module efficiency (from 6% in
1975) would be met as a result of the planned FSA
efforts. It was anticipated that this would occur as
solar cell efficiencies improved and as module
packing factors increased, i.e., more active solar cell
area per unit of module area. Logic indicated that
cell efficiency should increase as low-cost cell
design and fabrication technology matured, and as
mass production and quality assurance were introduced. Module packing factors of 0.90 were thought
to be achievable from the then 0.54 value. (1.0 equals
total coverage.) These presumptions have proven to
be accurate because, today, production modules fabricated using single-crystalline silicon cells have
efficiencies as high as about 11 % .
These efforts by many researchers and
technologists over many years have gradually
increased efficiencies from about 6 % when silicon
solar cells were devised in 1954 to the following
efficiencies at present:
• Production cells made from Cz ingots for
terrestrial use have efficiencies up to 14 %.
• Research cells made from Cz ingots have
reached 17 to 18%.
• Research cells made from float-zone silicon
have reached 19 % efficiency.
• Theoretical limit of crystalline silicon cells is
about 30%.
Economic analyses now indicate that module
efficiency must be about 15 % for photovoltaics to be
economically competitive for central-station
applications, which is now expected to be the first
widespread use of photovoltaics in the United States.
The achievement of 15 % efficiency modules requires
the repeatable fabrication of large-area laboratory
solar cells with 20 % or greater efficiency,
demonstrating that high-efficiency solar cell
principles are understood. This knowledge then must
be converted into low-cost production cell technology
that will yield uniform quality production cells of 17 to
18 % efficiency and, subsequently, module
efficiencies of 15 % .
• Probably maximum achievable limit is about
25%.
Project Activities: 1982 to Present
The general approach being used to increase solar
cell efficiency to greater than 20% has been
established. The required changes in certain cell
parameters to achieve this high efficiency are
believed to have been identified. However, the exact
'cell parametric changes and the cell design
modifications required so that large-area 20%
laboratory cells can consistently be fabricated have
not been devised. To assist in better understanding
how a cell functions and how various parts of a cell
influence its performance, the following figures and
description are included.
The objective is to perform research that identifies
and resolves key generic limitations to increased
efficiency of crystalline silicon solar cells as required
for large power generation applications.
The goal is to establish device technology required
for repeatable fabrication of large-area laboratory
crystalline silicon solar cells of greater than 20%
efficiency.
The main features of a crystalline silicon solar cell
are shown in Figure 22. The cell is made from a
semiconductor material doped (minute amounts of a
specific element have been introduced under
carefully controlled conditions) in a manner so as to
make the semiconductor sheet photosensitive. The
typical silicon cell made from p-type silicon (doped
with boron) sheet material is modified by the diffusion
or implantation of a phosphorus dopant to an
extremely thin layer (now n type) on the surface that
will become the top of the cell (emitter). The
semiconductors must be doped correctly, so that the
n-type material has large electron densities and the
p-type material has large hole densities. (A hole is a
vacancy in a silicon atom's outer electron ring after
an electron has been freed.) After the nip junction is
formed, a basic requirement for PV energy
conversion has been met by the creation of an
Background
Years ago, the need for high-power output from
spacecraft solar cells motivated a sustained research
for higher-efficiency cells. With the realization that
photovoltaics had the potential to become a practical
terrestrial power source, the research broadened.
Consequently, the performance of silicon solar cells
has been gradually increasing as a result of these
systematic efforts to gain an understanding of how
solar cells function, how to reduce cell losses by
better cell designs, and by improved processing. A
large number of relatively small advances, over many
years, have added to our knowledge of solar cell
technology and to increased cell efficiency. Early
38
160 . - - - - - - - - - - - - - - - - - - - - - - - ,
electrical field, which separates electrons and holes
within the cell. Near the junction, most of the
photogenerated holes from the n side transfer to the
p side; similarly, most of the photogenerated electrons
transfer to the n side. The nIp junction is permanent
and the electrical field is stable. When light shines on
a cell, the light that is not reflected is absorbed by
the semiconductors. When a photon enters the cell
and is of sufficient energy (appropriate wavelength)
and strikes one of the four outer (valence) electrons of
a silicon atom, that electron is dislodged. This leaves
a silicon bond missing an electron, called a hole, and
frees an electron to move about in the cell. Electrical current flow in a solar cell, when it does occur, is
due to both the motion of electrons and to effective
motion of the holes. When sunlight is absorbed in
the cell and frees electrons and holes from their
bonds, the electrical field separates the charges and
establishes an electrical potential, or voltage. If the
cell top and bottom contacts are connected to an
external circuit, an electrical current (amperage) will
flow as long as light continues to be absorbed by
the cell.
/I
MAX POWER
POINT
SC
I
120 f-
V OC
ISC
Vm
1m
Pm
FF
«
% EFF
E
;
;
;
;
;
;
;
640.9
143.5
539.8
134.2
72.4
0.787
18.10
mvolts
mamps
mvolts
mamps
mwatts
,.:
ffi
a:
a:
80 I-
::>
CELL AREA
o
INTENSITY
TEMP
4 SQ. CM
100 MW/SQ. CM
25 C
40 I-
I
I
I
I
i
I
o0L--L.--I-~_::'_0.2~-l--..L-.L--:0:'-:.4-.L.-....L-L.-...,.OL.6--l.L.-....J
VOLTS
Figure 23. Current-Voltage Curve of a High-Efficiency
Solar Cell
Sunlight
Front
Contact
~~~
!
MATERIAL, CHEMICAL AND ELECTRI·
CAL PROPERTIES
(THICKNESS, DEFECTS. GB, IMPURI·
TIES, RESISTIVITY, CARRIER LIFETIME, ETC)
GROWTH METHODS (CVDI, EVAPORA·
TION, OXIDATION, REFLECTANCE,
/~:::::;:=========:::::f"
Anti·Reflection
Coating
REFRACTIVE INDEX, LATTICE STRAIN,
SURFACE PASSIVATION (SiO x• Ta20S
SIN x)
DOPANT, DOPING METHODS, DOPING
CONCENTRATION, DOPING PROFILE
(GRADIENTS VS UNIFORMI, LOCALIZED
EFFECTS AT GBS, DEFECTS
nip Junction
~
~BUSSBAR
N+
Figure 22. Crystalline Silicon Solar Cell
AR COATING
TOP
EMITTER
JUNCTION
BASE
P-Si SHEET
For over a decade, there have been significant
improvements in crystalline silicon solar cell
performance, mainly as a result of increased cell
short-circuit current output. (Figure 23 is an I-V curve
showing the current and voltage of a typical highefficiency solar cell fabricated at JPL.) The
improvements have resulted by use of shallow
junctions, back surface fields, back surface
reflectors, textured surfaces, and multilayer
antireflection coatings. Shallow junctions help to
reduce the effects of cell front surface losses. Backsurface fields reduce the effects of back-surface
losses. Back-surface reflectors increase photon
absorption. Textured surfaces and multilayer
antireflection coatings reduce light reflection losses
at the front surface. Figure 24 shows a solar cell with
key factors that influence cell efficiency. The potential
for further improvements in a cell's short-circuit current are limited. However, single-junction crystalline
silicon cell performance can be increased by increasing the cell open circuit voltage. This is the area
of endeavor now being emphasized by FSA highefficiency silicon solar cell research.
t::~====~:2======BACK
SURFACE FIELD
c:
BACK METAL CONTACT
THICKNESS AND QUALITY OF BASE,
MINORITY CARRIER RECOMBINA·
TION, SURFACE PREPARATION,
SHARP VS GRADUAL (AI, B)
TOP METAL CONTACT, METALS (H
Pd·Ag; Ni, Cu, Ai, Ag), METHODS
CONCENTRATION DEPTH
CONTACT AND GRID RESISTANCE,
STABILITY WITH TEMPERATURE
(EVAPORATED, SCREEN PRINTED),
AREA COVERAGE, ME:TALS (TI·Pd·
Ag·, Ai, Ni, Ag, Cu) EVAPORATION,
SCREEN PRINTING SINTERING
Figure 24. Key Factors that Influence Cell Efficiency
The primary thrust of the FSA high-efficiency
efforts has been and continues to be directed
toward reducing solar cell bulk and surface losses,
with specific focus on reducing the minority carrier
recombination rates throughout the cell. Three
methods being used are: increased minority carrier
lifetime, reduced volume of silicon for bulk
recombination, and reduced surface recombination.
By obtai~ing silicon sheet with properties so that the
39
recombination rates in the bulk material do not
dominate cell performance, the emphasis can then
be on reducing recombination in the thin emitter
region and at the cell surface. Future cells will be
multilayer structures with each layer thickness
designed according to the electrical properties of
that layer and the other layers.
resistance, a short circuit is obtained. This causes
the pin junction region and the base and emitter
regions to discharge. If the discharge times related
to the junction and emitter regions are much smaller
than the time related to the base region, the lifetime
and recombination velocity can be determined.
Within FSA, the efforts being implemented are
directed toward gaining a better understanding of the
physics of carrier losses and their influence on cell
design by the use of comprehensive modeling, by the
development and use of new and improved
measurement techniques, and by new and improved
cell fabrication techniques. The surface
recombination velocity of a solar cell (a measure of
surface losses) is generally greater than 105 cmls
An effort is under way to reduce the surface
recombination velocity, to less than 103 cmls by·
addressing critical problems of controlling
recombination losses at surfaces and interfaces (for
example, the investigation of transparent conducting
materials). To better understand and control surfaceinterface phenomena, surface passivants (such as
SiN x , polycrystalline silicon, and ultra-thin oxide
layers) are under development. Also, reliable
measurement techniques to characterize passivants,
surfaces, and interfaces are evolving. As part of
surface loss investigations, work on transparent
conducting materials is emphasizing study of the
interaction of a polymer with a silicon surface and
the effects on optical and electrical characteristics.
Studies to understand and control bulk loss
mechanisms by improvements in silicon sheet
structural, electrical, and chemical properties are
leading to improved minority carrier diffusion lengths.
The complicated and varied nature of sheet is a
function of its growth conditions and its processing,
including solar cell fabrication.
R3
PULSE
GENERATOR
R1
LOG
AMPLIFIER
Figure 25. SCCD Cell Loss Mechanism Measurement
Method
One of the main advantages of this technique is
that the discharge time of the pin junction (on the
order of 10-11 s) is much less than the discharge
times associated with either the emitter or the base.
This makes it very easy to interpret the observed
transient current. The SCCD method is simple, fast,
and less expensive than methods used.to measure
other key solar cell parameters. It can be routinely
used during cell fabrication steps. It is being
evaluated as a possible method for determining
carrier lifetimes and recombination velocities in the
thin emitter region.
A highly sensitive microwave reflectance
measuring system with computer data acquisition
has been designed and built at JPL. The experimental
results illustrate that the m,icrowave technique can
probe accurately the decay of photogenerated
minority carriers in silicon sheet materials that
depend on the minority carrier lifetime and surface
recombination velocity. Under controlled surface
conditions, one can measure bulk minority carrier
lifetimes in silicon sheets, such as silicon wafers
grown by Czochralski and floating-zone techniques
and dendritic web ribbons. With known lifetime in the
bulk material, the technique has been demonstrated
to be an effective tool for monitoring the surface
recombination velocity. Preliminary results reveal that
the silicon surface polished by Syton has a small
surface recombination velocity when it is freshly
etched in a diluted hydrofluoric solution, but the
recombination velocity increases drastically when the
surface is exposed to the air. The effect becomes
saturated after several hours. Currently, JPL
scientists are working on an adequate analysis
method to extract quantitative results from
experimental data. The microwave technique is
reliable, nondestructive, effective, and easy to use. It
Characterization of bulk material and device
performance evaluation are using or will usea variety
of measurement techniques for chemical, structural,
and electrical characterization. These include optical
electron and ion beam analyses, X-ray characterization, light and dark current-voltage relationships,
spreading resistance, diffusion lengths by surface
photovoltage, scanning electron microscopy, deeplevel transient spectroscopy, secondary ion microprobe spectroscopy, and Zeeman spectroscopy.
Measurement techniques are being developed that
are critical to gaining a basic understanding of loss
mechanisms, to correlating model parameters with
measured ones, and to optimizing cell performance.
For example, the short-circuit current-decay (SCCD)
method (Figure 25) has been developed at the University of Florida. It simultaneously measures the
carrier lifetime in the bulk of the cell and the recombination velocity at the back surface interface. With
this method, a forward bias is applied to the solar
cell to set up a steady-state condition. Then, by
rapidly applying a zero bias across an extremely small
40
For future use, JPL is establishing improved
versions of the two models. Purdue's upgraded
model, SCAP2D, performs two-dimensional analysis.
This will enable researchers to analyze complicated
structural designs that are difficult to reduce to
one-dimensional equivalents. Stanford's new
SUPREM-1I1 also has additional features for
calculating electrical properties after each process
step.
can be used on as-grown and processed samples
without any contact.
Electron beam-induced current technique to
measure minority carrier lifetime in n + and surface
recombination velocity at n + front surface is being
investigated.
More complete analytical solar cell modeling and
analysis capabilities development has led to the
establishment of a program [Solar-Cell Efficiency
Estimation Methodology and Analysis (SEEMA)] that
uses computer models to evaluate and design
advanced silicon cells. After an extensive search,
FSA located two models to use in SEEMA. One of
these, Solar-Cell Analysis Program in One Dimension
(SCAP1 D), was designed by Purdue University. It is a
general purpose program used to analyze and design
silicon solar cells. The other SUPREM-II, was
designed by Stanford University. It is used to calculate information about the emitter doping profile which
is then used as input to the SCAP1 D model.
Recent Accomplishments
• New techniques to simultaneously measure bulk
minority carrier lifetime and back surface
recombination velocity developed, including SCCD
and microwave techniques.
• Key barriers to obtain high-efficiency cells
identified by use of comprehensive numerical
analysis techniques.
• Use of new passivants for front and back cell
surfaces showing promising results.
Researchers in FSA validated the utility of SCAP1 D
by evaluating the metal-insulator-nIp junction (MINP)
structure of the high-efficiency silicon cell reported
by Martin Green of Australia. Using experimental
data provided in the literature as input, SCAP1 D's
evaluation closely agreed with Green's experimental
results. Researchers also used SCAP1 D to perform a
sensitivity analysis on the structure by changing
various design parameters one at a time, while
holding the other parameters constant. Their analyses
indicated that a cell with the MINP structure could
obtain 20% efficiency by increasing the bulk minority
carrier lifetime from its present 20 ~m to 250 fLm.
This would make the diffusion length much greater
than the cell thickness of 280 mm (Figure 26). Under
these conditions, using a back surface field would
further improve the cell's efficiency.
• High-efficiency solar cells modeled.
• Theoretical analysis of a new cell design, called
"Floating Emitter Solar Cell Cell Transistor,"
indicates that the cell has the potential to achieve
greater than 20 % efficiency.
Research Needed
• Establish relationships between residue defects
and recombination mechanisms in highest quality
available silicon sheet.
• Investigate recombination mechanisms in low-cost
silicon sheet in bulk, in all layers, and at all
surfaces.
Lifetime (J.1 s)
• Develop technique for measurement of front
surface recombination velocity.
20
20
• Establish correlations between material defects,
device fabrication steps, and device performance.
19
• Improve modeling capabilities.
• Develop new higher-efficiency device concepts.
• Continue characterization of silicon sheets and
devices as their quality improves.
16
20
50 75 100
ED
200
Cell Thickness ( J.1 m)
Continue device modeling and experimental
verification of high-efficiency device structures.
• Investigate cell efficiency improvements for new
advanced cell processes.
Figure 26. Sensitivity of Efficiency to Solar Cell
Thickness for Various Minority-Carrier Lifetime Values
41
Process Research and Development
In 1975, the processes for fabrication of solar
cells, for use on Earth, were either those used for
making space solar-cells or their derivatives. These
techniques, borrowed from diode production
technology, were very labor intensive and material
intensive with respect to the total product price. Acid
etching and cleaning steps were extensively used
along with photolithographic grid-line definition.
The key technical issues are to develop processes
so that:
• Silicon sheet substrate quality is not degraded.
• A variety of high-efficiency cell designs can be
fabricated.
Background
The performance of individual solar cells was not
consistent. Fabricated in batches, even the
performances of those cells for use in space varied
significantly. The best of these small rectangular
"space cells" were assembled on panels, as closely
packed as possible, to produce the maximum power
per unit area. The lower performance cells were not
used in spacecraft power systems because of highperformance requirements. Instead, these reject
space cells were used for terrestrial applications. By
that time, solar cells made specifically for terrestrial
use had become competitive with reject space cells.
These circular cells were made from one complete
Cz wafer. The cell processing was similar, but
simpler and less expensive. For example, semiconductor industry-reject wafers were used. The
result was that terrestrial solar-cell performance was
less than space-cell performance. The challenge in
1975 was to reduce the cost of terrestrial solar cells,
but still retain good quality and performance.
Crystalline silicon solar cells were invented at Bell
Laboratories in the mid-1950s. These solar cells or
solar batteries, as they were called, are still the most
commonly used solar cells. A solar ,cell is fabricated
by a series of processes, from a silicon sheet doped
with either an n or p type impurity. The other
type of impurity (p or n) is then added to an
extremely thin (about 0.3 fLm) layer on the surface
that will be the top of the cell. This creates the nip
junction (or pin) where the photovoltaic effect can
occur in the solar cell. The addition of metallic
electric current collectors, on the top and on the
bottom of the wafer, permits electricity that is
generated within the cell to flow from the cell. The
top current collector consists of uniformly spaced
fine lines that minimize the amount of sunlight that is
blocked from entering and from being absorbed by
the solar cell. Cell fabrication, although it seems
simple, is sophisticated with improvements in
process and equipment continuing even after
30 years of progress. Completed solar cells are
electrically interconnected and then encased to form
a module that supports and protects the cells.
The objectives are to develop cell fabrication and
module assembly process technologies required to
meet cost and performance requirements for
terrestrial PV power production.
Project Activities: 1975 to 1981
Within FSA, the research and development
required for solar cell fabrication and module
assembly were conducted by a group familiar with
the processes and equipment used to produce solar
cells and semiconductor devices. A plan consisting
of five phases, which was compatible with the other
FSA activities, was devised and implemented. The
plan consisted of:
The goals were to:
• Identify and develop low-cost cell and module
production processes.
• Design facilities and equipment to perform the
processes.
• Demonstrate fabrication of low-cost PV cells
and modules in a pilot line environment.
• Implement the transfer of mass production
technology to industry.
I.
Technology Assessment: Determine availability
and applicability of cell and module processes.
II.
Process Development: Develop applicable old
and required new processes.
Facility and Equipment Design: Develop
automated equipment and facility designs.
III.
The goals now are to develop processes to:
• Fabricate high-efficiency cells from available
silicon sheet substrates.
IV.
Experimental Plant Construction: Incorporate
previous efforts into a pilot line to define yields
and remaining key cost factors.
• Assemble cells into higher-efficiency longer-life
modules.
V.
Conversion to Mass Production: Transfer
technology to industry.
42
The Technology Assessment contracts awarded to
Motorola, RCA and Texas Instruments in early 1976
assessed the state of the art of PV process
technologies in terms of cost large volume
production potential, ability to be automated,
reliability, and process synergistics. Identification of
those processes with the most potential for costeffective implementation and recommendations for
further study of those processes were also
performed. More than 50 processes covering all
areas of cell fabrication and module assembly were
studied. Processes evolved that clearly distinguished
terrestrial cells from space cells. As an example,
copper metallization was identified as being more
cost effective for top and bottom electrical
conductors than either solder or silver. Metallization
methods were expanded from space-oriented
vacuum evaporation to include electroless plating,
electrolytic plating, thick film metal (printing), and
reflow solder. Wafer surface treatments were
expanded to include brushing, plasma etching,
texture etching, and new cleaning solutions. Junction
formation, in addition to the standard space cell
process (gaseous diffusion), was expanded to include
spin-on and spray-on of liquid dopants, solid dopant
diffusion, doped oxides, ion implantation, and
advanced ion implantation. Other good candidate cell
processes that were added included silicon nitride
(CVD), oxide growths, mechanical edge grinding, and
laser scribing. Assembly methods included the use
of: glass substrates, glass superstrates, and
conductive adhesives; new encapSUlation materials
such as silicones, PVB, EVA; and new cell
interconnection methods. Selected processes were
studied regarding their "sensitivity to variables" such
as purity or condition of input materials and supplies
and processing parameters, primarily time and
temperature. Efforts at JPL were concentrated on
process verification and synergisms along with cell
metallization analysis programs.
" Solar Array Manufacturing Industry Simulation
(SAM IS), a factory simulation which runs on a
main frame computer. This document requires
the filling out of detailed costing sheets called
Format As which are then fed into a computer
for accurate and consistent process costs.
• Solar Array Manufacturing Industry Cost
Standards (SAMICS), a catalog of material
costs, labor costs, and other input costs which
assisted the contractors in inputting standard
costs.
These three techniques were verified and
upgraded continually throughout the life of the
Project and used extensively.
At the conclusion of the assessment effort, it was
determined that cost effective processes should:
" Consume a minimum of material and supplies.
" Use low-cost, recyclable supplies.
• Have high production yields.
" Result in high-efficiency modules.
Before the assessment effort, automation and
capital equipment were thought to be major cost
factors. It was shown that when labor is a significant
cost factor in a process, the process should be
automated. Most semiconductor processes were, to
a significant extent, already automated and,
generally, capital costs for equipment could be
distributed over large production runs resulting in low
costs per unit.
Requests for Process Development proposals
were distributed to the entire PV industry and were
broadly scoped to include new and novel processes
and those processes identified in the Technology
Assessment Phase as requiring further development.
Also, unsolicited proposals from PV contractors were
encouraged, reviewed, and acted upon. Twenty PV
industry contractors and two universities participated
in these studies, which reflected the thinking of the
PV industry at that time. A partial review of some of
the contractors and their diversified activities during
this phase are:
The problem facing the Project rather early in its
inception was how to compare the potential
production costs of competing processes and
process sequences being investigated by a number
of different researchers. This was clearly brought out
during the Technology Assessment phase in which
cost comparisons between the different contractors
on the same processes were significantly different.
This pinpointed a need for a standard methodology
that allowed relative comparisons of the potential
production costs attributable to competing
processes, and an estimate of the actual cost to
mass produce using a specific process.
" Applied Solar Energy Corp.: Low-cost copperplated contacts; ion-planted cell evaluation.
" Lockheed Missiles and Space Division:
Sprayed-on fluorocarbon encapsulation.
Three key costing analyses methods developed by
FSA PA&I personnel are:
" Mobil Solar Energy: Module assembly.
• Interim Price Estimation Guidelines (IPEG), a
costing system simple enough to be run on a
hand calculator. This method uses standard
coefficients for costing inputs such as
equipment, direct labor, material, floor space,
and utilities.
• Motorola: Antireflective (AR) coatings;
palladium/nickel/soldermetallization.
" Sensor Technology (later called Photowatt
International): Nickel/copper metallization.
43
Ell
RCA: Expitaxial cell processing.
Ell
Solarex: Wafer thickness studies; solar breeder
studies; high efficiency module; nickel/solder
metallization.
Ell
Sollos, Inc.: Nickel metallization.
Ell
Spectrolab: Thick-film metallization.
Ell
Spire Corp.: Ion implantation; pulsed electron
beam annealing.
Ell
University of Pennsylvania: Process ar:lalysis
and evaluation.
Ell
Westinghouse: Cell processing using dendritic
web.
spray-on dopant junction formation and printed metal
mask followed by plating; Spectrolab demonstrated a
spin-on dopant junction formation and a thick-film
metallization system; Motorola demonstrated ion
implantation junction forming, a CVD silicon nitride
AR coating and plasma etching; Solarex
demonstrated cost-effective processing of
polycrystalline material; Westinghouse indicated that
conventional semiconductor processes such as
gaseous diffusion and vacuum metallization could be
made cost effective with large volume production
equipment.
Collectively, the process sequences showed the
results of the cross-fertilization of technology that
FSA emphasized. Copper metallization was used in
two of the sequences. RCA cleaning technology was
used extensively, as was laser scribing and thick film
systems. The sequences illustrated the large number
of possible low-cost process combinations that could
be used to make a cell. The process sequences
demonstrated that there was a new breed of solar
cells, terrestrial solar cells, fabricated distinctly
different than the space cell from which it evolved.
All contracts during this phase required process
cost evaluations to be made using the costing
methodology and the writing of process
specifications. The costing documentation included
all of the direct inputs for equipment cost and
performance; required labor, floor space, utilities,
maintenance, supplies and input materials for each
process. The process specifications included detailed
descriptions of equipment, chemicals and materials
used, and process parameters including time,
temperature, pressure, etc.
Most of the process R&D efforts have been
conducted by private industry and universities
through subcontracts with FSA which, to date, have
included over 60 contracts. A parallel effort was the
creation of a Process Research Laboratory at JPL in
which the contractor process work could be verified
and, in many cases, directly supported. Independent
research has also been conducted in the laboratory
by JPL personnel.
Selection of an individual process depended upon
what other processes were used before and after
that process. Accordingly, an additional effort, the
development of process sequences, was initiated.
Discrete process steps were combined to form
process sequences that were synergistically sound
and cost effective, such as shown in Figure 27.
Contracts were awarded to develop and demonstrate
viable, cost-effective sequences. It became clear that
there was no one single process sequence that was
most cost effective, but rather a number of process
sequences could be demonstrated to be viable and
cost effective. Photowatt (then called Sensor
Technology) demonstrated a process sequence with
TYPICAL SEQUENCE 1977
A major tool used to transfer the evolving process
technology to the entire PV industry and to obtain
cross fertilization has been the distribution of process
specifications for verification. Near the end of the
Process Development Phase, more than 140 process
specifications covering cell and module fabrication
had been documented and distributed to interested
PV industry members. A feedback request resulted in
industry verification of more than 40 of these
processes. The results of these efforts are shown in
TYPICAL $2.80/WATT SEQUENCE*
TYPICAL $O.70/WATT SEQUENCE-
METALLURGICAL
SILICON
GROW INGOT
PHOTORESIST
'--------..----
SILICON PREP
GOLD PLATE
SOLDER COAT
CLEAN FLUX
SILICON PREP
'----...------'
SILICON PREP
~
SHEET FAa
'----.---'
CElL fAR
-------SHEET FAB
-------MOOULE FAB
'-------"..-'
MODULE FAB
Figure 27. Modu/e Process Sequences
44
"-----.--'
SHEET fAB
'---.----'
CELL fAa
"---..------'
MODULE FAD
effort by providing a place for supplementary
experimentation and verification.
Figure 28. The JPL Process Research Laboratory was
also very much involved in the process verification
An important function of FSA is to promote the
advancement of the photovoltaics industry by the transfer
of technology
~I
~
PIM
r-_S~~IREFf
I I
Former Technology Transfer
Procedures
circa 1981
-r
• VERIFY SELECTED
PROCESSES
• ASSESS SENSITIVITY
TO PROCESS VARIABLES
Iii
• VERIFY ECONOMIC
DATA fROM
CONTRACTOR
Processing Technology
Transfer Status
1981
Available
Surface Preparation
Junction Formation
Metallization
Module Assemblv
Totals
Under
Evaluation
Confirmed
19
18
13
16
7
4
6
5
5
5
3
0
66
22
13
Technology transfer continues but in a less
formal manner
RECENT TYPICAL TECHNOLOGY TRANSFER
• UNSUPPORTED ("CREDIT·CARD") MODULE LAMINATION
• MEL T REPLENISHMENT EQUIPMENT
Westinghouse Electric Corp. is using equipment
developed by Kayex Corp. and modified at Westinghouse
to replenish silicon in a dendritic·web growth process.
The first unsupported (no superstrate or substrate)
module was fabricated by Spectrolab using the JPL
Process Research Task's 4 x 4-ft laminator.
• CELCAL PROGRAM AVAILABLE FROM COSMIC
• CulnSe2 MODULE LAMINA TlON
Boeing Eng'g & Construction used the JPL Process
Research Task laboratory's equipment and processes
to laminate CulnSe2 cells.
• THERMALL Y INSULATED TOOLING FOR LAMINATORS
Sunelco Corp., Tideland Signal Corp., Solarex Inc. and
Boeing Eng'g & Construction have shown great interest in
thermal insulation techniques used in the JPL Process
Research Task's laminator (capable of producing 4 x 4·ft
modules) and are adopting them.
• TERRESTRIAL MODULE PRODUCTION LINE
Tideland has set up a production operation in Australia
based on processes transferred from Spectrolab, Inc.,
developed under contract with JPL.
An engineer-interactive computer program for solar·cell
metallization design has been transferred to and is
available from Computer Software Management and
Information Center (COSMIC).
• AUTOMATED ROLLING-SPOT BONDER CELL STRINGING
Kulicke and Solla, Inc., has built an automated machine
for Westinghouse based on a concept developed under
contract with JPL.
• THICK·FILM INK AVAILABILITY
Silver and aluminum thick·film inks are available from
Electrink, Inc., based on formulations developed under
JPL contracts.
Figure 28. Technology Transfer to Industry
45
In late 1979, a bill sponsored by Senator Tsongas
of Massachusetts provided additional funds for near·
term reduction of solar energy production costs by
soliciting ideas from industry. The following contracts
were implemented:
Organization
interconnect machines by ARCO Solar and Kulicke &
Soffa and a vacuum laminating system by ARCO
Solar. First·generation machines are shown in
Figure 29.
The Facilities and Equipment Design Phase was
never heavily stressed because most of this
technology was already being developed for non·PV
semiconductor devices. In addition to the special
near·term funding, the development of other
process equipment was funded. Important process
equipment developments were ion implantation
equipment for large area junction formation and a
pulsed electron beam wafer annealing unit by Spire
Corp., and a robotic cell interconnection and
module assembly system by Tracor MB Associates.
The robotic equipment was especially meaningful
because it demonstrated that cells could be handled
robotically in large·volume quantities without
breakage. These machines demonstrated the
potential for high·volume production.
Activities
ARCO Solar
Automated cell interconnection
Kulicke & Soffa
Automated cell interconnection
Motorola
Wax patterning.
Thin cell processing
Photowatt
Polysilicon cell processing;
surface texturizing
Sollos
Thick·film metallization
RCA
Megasonic cleaning
The important assembly equipment that evolved
from the near·term effort were the automated cell
AUTOMATED SOLAR CELL ELECTRICAL
INTERCONNECT (SERIES) ASSEMBLY
BY KULICKE AND SQFFA
1ST INTERCONNECT SOLDER
2ND INTERCONNECT SOLDER
EDGE SEAL AND FRAME ASSEMBLY
(TO BE DEVELOPED)
Figure 29. First-Generation Module Assembly
46
proposed module assembly technique for both
processes.
In late 1980, two Module Experimental Process
System Development Unit (MEPSDU) contracts
were awarded to provide the first controlled pilotline data on cell and module processing. They were
to reflect the culmination of the technology gained
from the earlier phases. Prior to this time, the
contractors produced cells and modules in a
laboratory environment and then extrapolated the
data for input into the SAMIS factory simulation
program. In the MEPSDU contracts, each of the two
contractors was required to construct and operate
a pilot line. The contractors were required to record
the production rate, yield, and process parameters.
Three technical demonstration runs were required,
two of which allowed no major adjustments,
modifications, or repairs during and between runs.
Solarex, Inc. and Westinghouse were competitively
selected for the MEPSDU contracts. The two
contractors chose diametrically opposite
approaches as shown in Figure 30. The Solarex
process sequence would produce lower efficiency
cells using low-cost processes and polycrystalline
silicon substrates. The Westinghouse process
sequence would result in higher-efficiency cells
using more expensive processes. The Solarex
process sequence used semicrystalline silicon
wafers, whereas the Westinghouse process
sequence used dendritic web silicon. Ultrasonic
bonding of cell interconnections to the solar cells
was to be used by Westinghouse and soldered
connections by Solarex. Vacuum lamination was the
However, within 6 months the MEPSDU contracts
were reduced in scope because of Project budget
reductions. Within a year and a half, the MEPSDU
efforts were cancelled entirely and the contractors
were redirected toward research-oriented efforts in
accordance with new Project guidelines.
Project Activities: 1981 to Present
The first three phases were completed, or almost
completed, as planned with significant reductions in
cell and module fabrication costs. Project
redirection starting in 1981 resulted in cancellation
of the MEPSDU and Mass Production Phases.
Good progress had been made toward meeting
the individual parts of the original Project goals.
However, an end-to-end integrated sequence was
never accomplished. Efforts were then directed
toward basic process research and, later, toward
higher-efficiency cell process research. With new
emphasis on higher efficiency and longer life,
development of more advanced and sophisticated
technology was initiated. Also, an examination of
the needs of the PV industry, at that time, and the
results of the earlier efforts led to the conclusion
that metallization and overall process yields were
two key items in which efforts should be continued.
APPROACH:
DESIGN AND DEVELOP EXPERIMENTAL
PROCESSING EQUIPMENT WHICH WILL HANDLE BOTH
FIRST AND SECOND GENERATION MATERIAL.
OBJECTIVE:
VERIFY PROCESS AND EQUIPMENT
CONCEPTS AND TRANSFER TECHNOLOGY TO INDUSTRY
r==r
c;
\ \ \ \ \ \~N
I
I
SI'R.AY ON
DOPANT
P
DIFFUSION
HEAT
F~
I
\\\\\\(
ALUMINUM PRINT
P
P+
SINTER
ANTI·REFLECTION (A·A)
t
~US~Cl~~~ ISO~n~~NA.R
II
~.N
P
P+
LASER SCRIBE Ni
JU~R
P'-
LASER SCRIBE
S~LDERM
f~~
Ni
CLEAN & SOLDER DIP
Cu
m'ft
MP0:l~:
NICKLE PLATE
Ni
A-R
EL
E~p.
COPPER PLATE
PR-METAL
REJECTION
[r:tS:lf
ETCH
A-R
P-
GASEOUS
DIFFUSION
MASK
P-
ETCH GLASS
MASK
DIP COAT
ANTI·REFLECTION IA·RI
~: r!
PRINT MASK
t::::ti: r
METALLIZEIVACUUM)
E~R
t~
L~:
P
GASEOUS P
DIFFUSION
:1~:
r
HOTO
RESiST (PfrRR
~_
II
EXPOSE-DEVElOp·ETCH
CELL STRINGING IULTRA-SONIC BOND)
Figure 30. Module Experimental Process System Development Unit
47
E=:+
SPRAY COAT
PR
E3
N
DIP COAT
The Westinghouse MEPSDU effort was redirected
to investigate liquid-dopant processing for junction
formation, which ultimately replaced the gaseous
diffusion process on the Westinghouse pilot line. The
Solarex MEPSDU effort was redirected toward
investigating process mechanisms affecting lifetime
and performance of cells made from polycrystalline
material. A second generation of automated cell
interconnection equipment evolved from the
Westinghouse MEPSDU effort. The initial effort was
cost-shared with JPL and, after MEPSDU
cancellation, was continued to completion by
Westinghouse. Kulicke and Soffa, under contract,
developed, designed, and fabricated the ultrasonic
bonding machine, which is capable of .
interconnecting one cell every 3.5 seconds or a 180
cell module every 10.5 minutes. The machine is
shown in Figures 31 and 32. Figure 33 shows interconnected dendritic web cells.
Figure 33. Ultrasonic Interconnect Bonding of
Dendritic Web Solar Cells
Metallization was determined to be the major
remaining higher-cost factor for cell processing.
Figure 34 shows the results of cost analysis efforts
on metallization systems. A number of innovative
metallization ideas were generated as a result of the
Cell Metallization Research Forum held in 1983. The
objectives of the Forum were to clarify and define
the state of the art of current solar cell metallization
systems to: report on performance experience with
these systems, including advancements made and
problems encountered; describe advanced
processing techniques under development; and to
discuss expectations for future improvements.
Figure 31. Second-Generation Cell Stringing Machine
Process control became a major area of concern
as industry yields were not as high nor as predictable
as had been expected. Industry proprietary concerns
and lack of large-scale engineering pilot lines limited
the availability of data. To overcome the lack of
information and to explore processing for higherefficiency cells, process and device modeling
capabilities were investigated. For example, SUPREM
(developed for the semiconductor industry by
Stanford University) provides a comprehensive
computer process simulation program that allows the
prediction of device structures from any fabrication
technology sequence. An example of a device model
is SPCOLAY, by the University of Pennsylvania, a
one-dimensional PV cell model which has been
particularly useful in defining the interactions
between cell substrate material parameters and
preferred cell processes. The soft boundary shown in
Figure 35 defines the "Domain of Practical Significance" for those characteristics which are prevalent
in present state-of-the-art materials.
"\
MOLECULAR BOND.\
Figure 32. Rolling Spot Ultrasonic Bonding
48
24.--,.---,---,---,--,,.--,----.,.-----,
BULK RESISTIVITY, D.·em
BSF THICKNESS, f..Lm
Candidate Processes/Systems
COSTlm 2
PROCESS/SYSTEM
10
NO BSF
OATA SOURCE
Evaporation (EV)
22
• SOA (TiIPd/Ag)
• Advanced lTilNi .-to Cu plating')
$90
$14
ASEC
Westinghouse
~10
::.:::::::;;;;;;;;--~ 5
"ff~&iJ#oP"{
1
0.1
NO BSF
Screen print (PRINT)
•
•
•
•
•
SOA (Ag paste)
1g90 (Ag paste')
SOA (AI paste)
1g90 (AI paste')
1990 (Mo/Sn')
$8 to $35
$7 to $34
$2.4
$1.4
2.80/W, Block IV"
JPl BPU"
2.80/W, Block IV
JPl BPU
$2.0
JPl BPU, Sol/los
10
1
0.1
>-"
~
20
LU
U
0.01
u::
u..
NO BSF
LU
Electroless plating
Z
• SOA (Print resist, Ni·plate, Sinter,
Wave solder)
$8
• Advanced (PR, Ni plate, Sinter, Cu
plating')
$9.4
Midlilm' (Ag) (MID)
Midfilm' (MoISn)
Ion plating' (Ti/Ni/Cu) (ION)
Cl
c;:;
Solarex, Motorola
a:
LU
~
Motorola
18
Cl
u
$5 to $16
$2
$6
Spectrolah"
Spectrolab, Sol/los
Illinois Tool Works
.. Advancement of Photovoltaic state-of-the·art
"Ag price from $10 to $5010'
"DOMAIN OF
PRACTICAL
SIGNIFICANCE"
Current Results
400
Figure 35. Low- and High-Bulk Resistivity Cell
Behavior; Effects of Bulk Resistivity, Diffusion;
Length, and Back Surface Field
PRINT RESIST
ELECTROLESS
NilSOLOER 0
The term "directed energy research" was coined
to describe the increased emphasis on highefficiency processes, which covers the use of
microwave, laser, and thermal pulse energy sources.
All of these new energy sources are of technical
interest because they allow many cell processes to
be performed without heating the bulk of the
substrate. PV cell fabrication by "cold" processing is
used to avoid degradation of bulk substrate
properties, particularly minority carrier lifetime.
FRONT METALLIZATION COST. $lm 2
PURPOSE OF THIS ANAL YS/S:
• Compare costs and effectiveness of state'
of-the'art (SOA) metallization and projected
metallization approaches
• Estimate the potential impact of R&D in this
area
Results from directed energy research have been
very gratifying. Westinghouse demonstrated that
thermal pulse (heat lamps) can be used to
simultaneously form nand p + junctions using liquid
dopants. Previous processing required more
handling, cleaning, and the use of expensive oxide
masks. Excimer lasers can be scaled up and have
been used by ARGO Solar and Spire for annealing
ion-implanted junctions. Superwave Technology
developed a microwave-powered plasma system
used for thin-filmed silicon nitride deposition as well
as a passivating silicon dioxide layer. Argon lasers
were used by Westinghouse in a successful new
metallization method in which lasers scan wafers
covered with metal, bonding the metal to the solar
cell where the "laser writing" has occurred,
eliminating the costly photolithography process
(Figure 36).
APPROACH:
• Use data from FSA contractors as the major
source for calculating cell metallization
costs with IPEG2
• Use JPL's grid optimization model to
establish electrical performance ratios
based on geometrical and electrical data
(ratio to a cell without resistive or shadow
losses)
• Develop metallization cost vs performance
trade'off slopes for several PV systems; the
cell metallization system farthest above a
trade·off slope has the greatest value
• Study as yet has not included process
compatibility, reliability, or unconventional
metallization systems
Figure 34. Metallization Cost Comparison
49
J :4==== Ar+
SCANNING MIRRORS
LASER BEAM
"""=::::::1t=:::::=;;l>
LEN S
SUBSTITUENTS
DEPOSITED
METAL FILM
R1
R2
I--
R3
R1
R2
SILICON
______ SUBSTRATE
R3
R1
R2
R3
SAMPLE BASE TEMPERATURE 75°C
FOCUSED LASER SPOT DECOMPOSES SPUN-ON FILM
METALLIZATION PATTERNS ARE FORMED BY DIRECT WRITING
Figure 36. Laser Pyrolysis of Spun-On MetalloOrganic Film
% ABUNDANCE
CH3
CH3
C6 H 13
= CH3
> CH3
< CH6H13
> CH3
> CH3
< C5 H l1
31
67
2
Figure 38. Alternative Formations of Metallo-Organic
Deposition Precursor Chemicals
Today's Status
One of the new metallization research efforts
funded was the development of metallo-organic
decomposition (MOD) films by Purdue University. The
MOD process uses liquid compounds containing
metal atoms bonded to organic compounds such as
neodecanoic acid. Figures 37 and 38 illustrate the
wide variety of possible compounds that can be
synthesized for research or for use in production.
Decomposition of these compounds by the
application of heat between 250 °C-300 °C results in
the formation of fugitive gases such as carbon
dioxide and deposition of a thin layer of metal. Many
different metals have been successfully tried such as
titanium, chromium, silver, bismuth, and palladium.
Combinations of metals have also been successful
with especially good results obtained from a bismuthsilver formulation. Choice of organic acid and various
viscosity adjusters allows application of MOD inks by
spinning, dipping, spraying, screen printing, and ink
jet printing. Multiple layers have also been successfully deposited, and lines as narrow as 3 fLm have
been achieved.
Low-cost processes have been developed for cell
and module fabrication steps such as:
• Surface Preparation
.. Hot alkaline damage etch.
.. Spray, dip, roller, spin, and meniscus coating
of anti reflective films.
• Junction Formation
.. Gaseous diffusion of 6-in.-diameter cells.
.. Liquid-dopant junctions.
.. Liquid-dopant back surface fields.
.. Aluminum-paste back surface fields.
• Metallization
HETERO-ATOM
(0, S, N, P, As)
.. Silver thick-film ink systems.
+
I
=
=
=
I
.. Copper cell metallization systems.
COMMON SOLVENT
+
.. Pioneering of molybdenum-tin inks.
I RHEOLOGY ADJUSTORS I
II
• Module Assembly
~
.. Ultrasonic cell interconnection.
+
PRINT, SPRAY, DIP, ETC
I RAW
FILM
I
.. Automated cell stringing.
+
.. EVA encapsulation processes.
I HEAT I
II
.. Double-chamber vacuum lamination.
ICONDUCTOR I
The above processes seem to have the potential
to meet the original Project goals.
Figure 37. Metallo-Organic Deposition Processing
50
• Laser annealing, electron pulse beam
annealing, and rapid thermal pulse annealing
of ion implanted junctions as well as
conventional thermal annealing to maintain
lifetime of the initial silicon material (Spire,
ARCO Solar, Lockheed).
A complete processing sequence using all the
steps, the individual yields, and the process parametric sensitivities was not completed.
The need for higher-efficiency and longer-life
modules, but manufactured at essentially the same
costs has imposed the need for new, more advanced
technology and processes.
• Spin-on, spray-on, and meniscus coating of
liquid dopants (Spectrolab, Sensor
Technology, Westinghouse).
Accomplishments
The following is a list of accomplishments with
the associated contractor(s) credited.
• Laser scribe (Sensor Technology).
Processes
ED
ED
Surface Preparation
Metallization
e
• A cost-effective etching processing for saw
damage removal prior to junction formation
(Sensor Technology, Lockheed).
• Metallo-Organic decomposition films (MOD
films) using low-temperature processes
« 300°C) (Purdue, Electrink, Westinghouse).
• Anistropic etching process to produce large
area textured surfaces for enhanced photon
collection (Motorola, Sensor Technology,
Solar Technology).
• Reliable plating systems using Pd and Ni
followed by solder build-up by immersion
(Motorola, Solarex).
• Developed a number of antireflection
coatings with concurrent processes,
including mUlti-layer evaporation, spraying,
dipping and spinning, and the required
material development (RCA, Westinghouse).
• Pyrolytic decomposition of MOD films using a
laser as the thermal source (Westinghouse).
e
• Polycrystalline silicon processing (Solarex).
• Microwave enhanced CVD (Superwave).
ED
Megasonic cleaning (RCA).
e
Plasma pattern etching (Motorola).
ED
e
• Automated soldering of interconnects (ARCO,
Kulicke & Soffa, Foothill Engineering).
• Laminated encapsulation in a partial vacuum.
Large area, large volume, gaseous diffusion
processing using POCl s, and PHs and BCl s
(Sensor Technology, Westinghouse).
e
Simultaneous front and back junction
formation using liquid dopants and infrared
(IR) rapid heating pulse techniques
(Westinghouse).
• Ion implantation of both front and back
junction at a rate of 1200 wafers/hour (Spire).
e
Module Assembly
• Ultrasonic bonding of interconnects
(Westinghouse, Kulicke & Soffa).
Junction Formation
e
Development of a generic fabrication
process for producing MOD film (Purdue,
Electrink).
• Diffusion barriers (Caltech).
• Surface passivation of front emitter areas
using silicon dioxide, silicon nitride, and
silicon oxynitride (Motorola, Westinghouse,
Spire).
e
Thick film screenable cost-effective
metallization processes using Ag, AI, AgAI,
Cu, and MOD AgBi (Spectrolab, Bernd Ross
Assoc., RCA, Purdue, Electrink, Sollos, Solec).
Non-mass analyzed ion implantation of both
back and front junctions (Motorola, Spire).
PVB and attendant processes for 4 x 4 ft
module (ARCO Solar, RCA).
e
EVA and attendant processes (RCA).
e
PV shingle assembly (General Electric).
e
Silane primers, antioxidants, and UV screening agents (Springborn Laboratories)
• Electrostatic/ultrasonic edge sealing (Spire)
(see Figure 39).
51
ALUMINUM TO
ALUMINUM SEAL
BY ULTRASONiC
• Automated cell stringing machine (ARCO
Solar, Kulicke & Soffa).
ENCAPSULANT
WELD
\
/
ALUMINUM TO
GLASS SEAL
BY ES8
• Ultras.onic automated cell stringing for
dendritic web (Westinghouse, Kulicke &
Soffa).
SOLAR
CELL
ANTICIPATED BENEFITS
APPLICATIONS
III REDUCED H20 VAPOR
TRANSMISSION
III MODULES IN SEVERE ENVIRONMENTS
III REDUCED 02 TRANSMISSION
III REDUCED AIRBORNE
POLLUTANT TRANSMISSION
GI UV AND TEMPERATURE
STABLE SEAL
• A laminator capable of producing large area
modules (JPL, ARCO Solar, Spire, ASEC).
GLASS
SUPERSTRATE
" Robotic module assembly (Tracor
MB Associates).
- COASTAL AND MARINE SITES
- HOT AND HUMID REGIONS
GI REMOTE LOCATIONS
Analyses
- REDUCED MAINTENANCE
AND REPLACEMENT COSTS
GI MODULES WITH SENSITIVE
COMPONENTS
Ell
SPCOLAY (University of Pennsylvania). PV cell
modelling program for investigation of process
variable sensitivities.
Ell
Zero depth concentrator effect (JPL, Science
Applications, J. Mark, General Electric).
Mathematical model of the effect of light
trapping by non-specular reflection in thick
films.
Ell
Grid line optimization (JPL, Texas Instruments,
Motorola, RCA, Spectrolab). Mathematical
models and computer optimization programs to
allow design of minimum power loss collector
grids.
Ell
$2.00/watt (JPL). "Strawman" process analyses
for low cost cell and module fabrication.
" Non-mass analyzed ion implant equipment
(JPL, Motorola, Spire).
Ell
$0.50/watt (JPL). "Strawman" process analyses
using projected, automated processes.
" Pulsed electron beam annealing system (Spire).
Ell
Diffusion barriers (Caltech). Development of the
scientific basis for formation of diffusion
barriers, especially those of amorphous films.
Ell
Solar breeder (Solarex). An analysis of a
photovoltaics production facility powered by
photovoltaics.
Ell
Wafer thickness (Motorola, Solarex). An analysis
of wafer thickness cost tradeoffs.
Ell
Polycrystalline cell processing '(Solarex, Sensor
Technology). Development of techniques for
processing polycrystalline silicon material.
Ell
Solar Area Production System Handbook
(Theodore Barry Associates). A compilation
of costing variables related to large volume
production of solar cells.
Ell
Tandem Junction cell (Texas Instruments). An
analysis of a novel photovoltaic device structure
to produce electrical power.
GI EXTENDED LIFETIME FOR
STANDARD MODULES
Figure 39. Hermetic Edge Sealing
Equipment
Ell
Surface Preparation
• Megasonic cleaning equipment (RCA).
" Microwave enhanced CVD System
(Superwave).
Ell
Junction Formation
" Ion implanter with 4-mA capability for
phosphorous and throughput of 400 wafers/h
(Spire).
• Laser annealing system (Spire, ARCO Solar,
Lockheed).
• Meniscus coating machine for dopant
additions to dendritic web material
(Westinghouse).
Ell
Metallization
• Ink jet printer-computer controlled (Purdue).
• Argon ion laser decomposition system
(Westinghouse).
" Dry film/mid film (Spectrolab, Applied Solar
Energy Corp.).
Ell
Module Assembly
" Anti-reflectance treatment of glass
(Motorola).
52
Modules and Arrays
• Verify adaptability of technology to meet
required operational conditions.
In 1975, crystalline-silicon solar cells for use on
spacecraft had already been manufactured for about
15 years. There was significant progress made
during these years in understanding how solar cells
functioned and how to improve their performance.
However, there was little information about how to
design, fabricate, and use PV modules and arrays for
terrestrial applications. There was an immediate
need for module and array knowledge and
experience. A complete technology had to be
established including environmental and design
specifications, engineering concepts and design
knowledge,encapsulation materials and processing,
module assembly techniques, quality assurance
techniques, module performance and durability
evaluation techniques and facilities, and failure
analysis techniques and capabilities. An intensive
effort to meet this need was initiated with the start of
the FSA Project in 1975.
• Transfer the above technology to industry.
The goals, as established in 1983, are to develop
flat-plate (non-concentrating) module and array
technology as follows:
• Modules with crystalline silicon solar cells.
• High efficiency, 15 %.
• Low cost, $90/m 2 (1982 dollars).
• Thirty-year reliable operation.
• Modules with thin-film solar cells.
• Efficiency, 12%.
• Low cost, $70/m 2 (1982 dollars).
The objectives of the module and array activities
for both crystalline silicon and thin-film technologies
are to:
• Thirty-year reliable operation.
In 1975, the goals established were to develop flatplate module and array technology that industry
could use as early as 1986 to achieve:
• Develop design requirements and specifications
compatible with various operational and safety
environments for a variety of applications.
• Module prices less than $0.50 per peak watt
FOB (1975 dollars), marketing and distribution
costs not included.
• Develop the engineering sciences technology
required to integrate low-eost, efficient cell
technologies into cost effective and safe
modules and arrays that meet the operational
requirements of future large-scale applications.
• Production rates sufficient to attain economies
of scale.
• Develop the technologies required to achieve
and to verify reliable 30-year life flat-plate PV
modules.
• Operating lifetimes in excess of 20 years.
• Module efficiency greater than 10%.
53
Background
and error. The products had high failure rates, caused
by both immature designs and poor processing.
Early in the 1970s, small modules were made
using reject space cells sealed into heavy cast-glass
modules that were durable, but expensive (Figure 40).
However, typical mid-1970s terrestrial modules,
designed for more economical manufacturing, had a
lifetime ranging from 6 months to 2 or 3 years (Figure 41). The solar cells for these modules were made
using circular wafers directly sliced from a Cz ingot
and processed by essentially space-cell technology.
The electrically interconnected cells were usually
placed on a layer of silicone rubber covering a fiberglass panel, covered with a second layer of poured
silicon,8' rubber and cured. The' entirel process!was!very
labor intensive and subject to product variations. The
module designs and processing were often by trial
During the past 10 years, the FSA-industry
partnership has significantly advanced the basic
technology for terrestrial module and array design
and fabrication, using crystalline-silicon solar cells.
This was accomplished by concurrent experimental
and analytical research efforts supported by iterative
fabrication and testing of developmental and
commercial modules, as described in Module and
Array Evolution, Engineering Sciences and Design,
Encapsulation, Reliability Physics, and Module
Performance and Failure Analysis. The evaluation of
modules using thin-film solar cells is just beginning,
as is the adoption of existing knowledge and the
development of unique new techniques for thin-film
modules.
Figure 40. Durable Modules Made from Space Solar
Cells in Early 1970s.
54
Figure 41. Mid-1970s Terrestrial Photovoltaic Modules
Project Activities: 1975 to 1981
and Government. This effort resulted in a continuing
iterative evolution of new knowledge and capabilities.
Practical, realistic environmental protection
requirements and approaches were developed along
with techniques for evaluating module capabilities for
meeting these specifications. Practical module
design knowledge grew as encapsulation materials
and processing methods were developed and
evaluated, and the compatibility of the modules with
PV power systems was established. As the
technology evolved, it was incorporated by industry
into durable hardware. This was accomplished by the
Project sponsoring cell and module technology
Initially, in 1975, contemporary modules were
purchased and evaluated; concurrently,
developmental work was initiated on module
specifications, designs, materials, and processing.
These studies evaluated both the technical and
economical feasibilities of various technical
concepts. During the next few years, simultaneous
assessment of the needs, capabilities, deficiencies,
and potential solutions across the entire spectrum of
module technologies was accomplished by an
excellent cooperative effort by industry, universities,
55
developments, the purchase of advanced modules,
the evaluation of these modules, and the analysis of
module problems and failures (Figure 42). A flow of
information from operational PV systems, as shown in
Figure 43, helps the technology development activities shown in Figure 42. Five sequential purchases
("block buys" of modules designed and fabricated by
industry to JPL specifications) have resulted in an
effective cooperative effort by the FSA Project and
industry. The large quantities of modules purchased
-
(large at the time) for the first three block buys were
intended to increase industry production capacity as
well as provide modules for evaluation and field testing. Block IV and V modules were purchased in
limited quantities in the 1980s, for test and evaluation
only. These Project efforts were instrumental in
developing practical flat-plate module design and
manufacturing technology and in stimulating industry
to incorporate much of it into its commercial products.
RELATED
TECHNOLOGIES
RESEARCH
t
MODULE AND
ARRAY DESIGN ~
REQUIREMENTS
DESIGN SYNTHESIS
MODULE EVALUATION
ANALYSES,
EXPERIMENTS,
AND COST
EFFECTIVE SOLUTIONS
PROTOTYPE MODULE
QUALIFICATION AND
ACCELERATED TESTING
IN-FiElD EVALUATION
t
-
FAILURE ANALYSIS
DURABILITY MODELING
LIFE PREDICTION
AND ACCELERATION
TEST CAPABILITIES
~
QUANTIFICATION OF
DEGRADATION
MECHANISM
PARAMETERS
1..0.-
PRODUCTION MODULES
AND ARRAYS
LONG·TERM
OPERATIONAL
CONDITIONS
30-YEAR
LIFE
TECHNOLOGY
r-
The design requirements were periodically updated based upon new knowledge of application needs, environmenta-I
conditions, realism of evaluation and testing techniques and hardware performance.
Figure 42. Module/Array Technology Development
INFORMATION FLOW
FIELD SITES AND
APPLICATIONS
•
•
•
•
SYSTEM NEEDS/SENSITIVITIES
ARRAY FIELD EXPERIENCE
TECHNOLOGY CAPABILITY
RAW WEATHER DATA
FSA
ENGINEERING
•
•
•
•
ENGINEERING ANALYSIS
ENVIRONMENTAL ASSESSMENT
EXPLORATORY TESTING
DESIGN OPTIMIZATION
INDUSTRY
FSA PROJECT
SYSTEMS ENGINEERING
TESTS AND APPLICATIONS
STANDARDS ACTIVITIES
•
•
•
•
•
•
•
ENVIRONMENTAL STRESS LEVELS
PERFORMANCE CRITERIA
PERFORMANCE DATA
DESIGN GUIDELINES
LOW-COST DESIGNS
ANALYSIS TOOLS
TEST METHODS
Figure 43. Information and Flow
56
Project Activities: 1981 to Present
• Modules of 0.9 MW sold for just under $5 per
peak watt (1983 SMUD purchase).
Since 1981, the evolution of module technology
and the adoption by industry has continued with
emphasis now on high-efficiency, cost-effective
modules that meet safety and reliability requirements
of future large-scale applications such as for utilities.
Efforts on rooftop and residential applications are
fewer now that it is anticipated that utility
applications will be the first large-scale use of
photovoltaics in the United States.
• Estimated module price would be about $1.47
per peak watt if a multi-tens of megawatts
production plant were built with the latest
existing technology (see PA&I section of report).
Thin-Film Modules
Prototype modules (other than consumer products)
are only beginning to become available for
evaluation:
A second parallel effort, just starting, is
establishing the practicality of modules made using
thin-film solar cells. This requires going through many
of the above sequences using as much of the
existing knowledge and capabilities as possible;
however, many unique new techniques and new
knowledge will have to be developed and evaluated.
As an example, new accurate, repeatable electrical
performance measurement techniques for thin-film
cells and modules are being developed and verified
so that degradation can be accurately quantified and
performance measurements standardized. With the
expanding availability of first-generation thin-film
modules, performance evaluation and assessment of
reliability characteristics are under way.
• Reliability is unknown.
• Efficiency is about 3 to 5 %.
• Prices are about the same as for crystallinesilicon modules.
Thin-film module differences requiring new or
expanded research:
• New cell environmental durability
(temperature/humidity/ultraviolet) failure modes.
Today's Status
• Altered hot-spot heating failure mechanisms.
Crystalline Silicon Modules
• Short-circuit cell failure modes and effect on
size and series/parallel redundancy.
• Estimated life of contemporary production
modules is more than 10 years.
• New cell electrical-interconnect failure modes.
• Module efficiency is as high as +11% in
standard production modules.
• Altered glass breaking strength.
• Flexible substrate technology demands.
• Encapsulation materials have been life-tested to
20 years.
• High-cell stresses caused by glass bending.
• Comprehensive terrestrial module design technology and knowledge exist, which have been
gained from 10 years of analytical, design,
fabrication, and test activities in active
cooperation with industry.
• Non-linear electrical response and effect on
measurement.
• Cell-to-cell electrical variability and effect of
electrical mismatch on circuit design.
57
odule and
Array Evolution
evaluations were performed by FSA. These activities
served three functions: (1) they encouraged industry to
incorporate the latest technology developed in the PV
program, (2) the various modules were evaluated by
uniform repeatable tests that enabled identification of
design and/or fabrication deficiencies, and (3) they
enabled a cooperative manufacturer/evaluator quick
response to module problems that occurred.
The PV terrestrial modules have evolved during the
past 10 years so that they are now practical, reliable,
and cost effective for many applications. There have
been major improvements in design, fabrication,
performance, and reliability. These improvements have
resulted in large reductions in module prices and lifecycle PV power generation costs. An important factor
contributing to this evolution has been the series of five
module procurements known as "block buys." The
module design and development were performed by
industry to FSA specifications; the module performance
A module from each of the five block buys is
shown in Figure 44. Table 17 lists the representative
characteristics of each block of modules. The photograph, the list of characteristics, and the five trend
charts (Figures 45 through 49) portray the progress of
flat-plate PV modules with crystalline silicon solar
cells during the past 10 years.
IV
58 Walls'
!
5 Watts'
21
SOlAREX
SOLAR
POWER
GENERAL
mCTRlC
Figure 44. Examples of Block I Through V Modules
58
Table 17. Representative Characteristics of Block Modules
AREA Im 21
WEIGHT Ikgl
SUPERSTRATE OR TOP COVER
SUBSTRATE OR BOTTOM COVER
FRAME
CONNECTIONS
ENCAPSULATION SYSTEM
ENCAPSULATION MATERIAL
CELLS
QUANTITY
SIZE (mm)
CONFIGURATION
MATERIAL
JUNCTION
FAULT TOLERANCE
PARALLEL CELL STRINGS
INTERCONNECT REDUNDANCY
BY·PASS DIDOES
PACKING FACTOR
NDCT"
PERFORMANCE AT 2B·C CELL TEMP. b
POWER, MAX. IWI
MODULE EFFICIENCY 1%1
ENCAPSULATED CELL EFFICIENCY (%1
I
II
III
IV
V
0.1
2
SILICONE RUBBER
RIGIO PAN
NO
TERMINALS
CAST
SILICONE RUBBER
0.4
5
SILICONE RUBBER
RIGIO PAN
YES
J·BOX
CAST
SILICONE RUBBER
0.3
5
SILICONE RUBBER
RIGIO PAN
YES
TERMINALS
CAST
SILICONE RUBBER
0.6
9
GLASS
FLEXIBLE SHEET
YES
PIGTAILS
LAMINATED
PVB
1.1
17
GLASS
FLEXIBLE LAMINATE
NO
PLUG·IN
LAMINATEO
EVA
21
DlA: 76
ROUNO
CZ
NIP
42
DlA: 76
ROUNO
CZ
NIP
43
DlA: 76
ROUND
CZ
NIP
75
95 x 95
SHAPED
CZ
NIP p+
117
100 x 100
SHAPED
CZ
NIP
NONE
NONE
NO
0.54
43
NONE
MINOR
NO
0.60
44
NONE
MINOR
NO
0.65
48
3
MUCH
YES
0.7B
4B
6
MUCH
YES
0.89
48
B
5.8
10.6
24
6.7
11.2
26
7.4
11.5
54
9.1
11.8
112
10.6
12.3
I
"NOMINAL OPERATING CELL TEMPERATURE: CELL TEMPERATURE IN OPEN·CIRCUITED MODULE EXPOSED TO 80 mW/cm 2 INSOLATION IN AMBIENT OF 20·C, 1 m/! WIND VELOCITY.
bAT 100 mW/cm 2, AM 1.5 INSOLATION.
I
II
40 I-
I
I
I
I
I
-
CELL TEMP: 25·C
I
30 I-
I
I
12.5
"'-
INSOLATION: 100 mW/cm 2, AM1.5
>-'
-
u
i:E
II
I
l-
I
I ESTIMATED
I
76
I
I
78
77
V-
c
c
:;;
I
I
79
80
III
~
7.5
I
5.0
I
::::>
I
75
~
T
I
I
-
IV
III
10 l-
I
10.0
c:;
I
I
82
81
I
83
I
II
I
IV
V
I
I
I
84
CELL TEMP: 25'C
INSOLATION: 100 mWlcm 2, AM1.5
75
82
84
YEAR
Figure 46. Module Efficiency Trend
Figure 45. Module Cost Trend
200
;:
a:'
100
~
a:>
~
~
::::>
c
c
:;;
III
II
I I
20
10
"'- 15.0
,.::
I
i:E
c:;
~
12.5
u
10.0
~
~
~
53
76
77
78
I
I
I
79
80
81
I
7.5
'"e<u
INSOLATION: 100 mWlcm 2, AM1.5
III
II
I
I-
:5
::::>
CELL TEMP: 25·C
<l-
i:E
75
IV
u
VI
IV
5.0
84
I
I
I
I
I
I
I
J
IV
cr
c
-
II
U
;:;:
'";;:2:
0.5 i -
I
I
II
I
I-
-
u
e<
<I-
-
-
o'----_'--1_1'----_'--1_'--1_'--1_-'--1_-'-I-'-I-:-:-~I---::-:-'
75
76
77
78
79
80
81
82
YEAR
Figure 49. Packing Factor Trend
59
CELL TEMP: 25'C
84
75
Figure 48. Cell Efficiency Trend
1.0 -
[
INSOLATION: 100 mWlcm 2, AM1.5
YEAR
Figure 47. Module Power Trend
V
83
84
The following sequence of events typifies the
activities of a recent block buy:
(II
(II
needs of the users and the desires of the
manufacturers.
FSA prepares design and test specification based
upon prior experience.
(II
FSA conducts competitive procurement
culminating in award of parallel contracts for
alternate design concepts.
(II
(II
Contractor performs module design.
(II
FSA conducts design review.
(II
Contractor fabricates 10 prototype modules.
(II
(II
FSA performs module qualification tests (and
failure analysis, as applicable).
FSA conducts design review.
(II
Contractor fabricates 10 modules.
(II
(II
(II
Contractor modifies design and/or processing
procedures to correct problems revealed by
qualification tests.
(II
(II
(II
FSA performs module qualification test (and
failure analysis, as applicable).
(II
Contractor modifies design and/or processing as
necessary and supplies modules for retest.
FSA prepares and issues User Handbook
describing construction details and performance
of successful module design by each contractor.
This series of module procurements has produced
the following results in PV module technology:
Ell
Adoption by the international PV community of
the FSA-developed specifications and design
features of the Block V module.
Variations of three of the Block V modules
being purchased for the second phase of the
1-MW PV central station power plant being
The competitive nature of the module
procurements which included updated design
and test specifications and a work statement
indicating design parameters requiring
improvement.
FSA completes final testing.
(II
(II
Systematic development of module failure
analysis techniques, equipment and
supplementary analysis, and experimentation
leading to the solution of module problems.
The success of these activities can be attributed, at
least in part, to the following:
(II
(II
Development and continuing improvement in
module evaluation capabilities and equipment
including electrical performance, physical
qualification tests, and correlation of both with
field experience.
A systematic upgrading of module quality and
performance by the incorporation of new
technology into improved designs that are more
amenable to mass production.
(II
Extensive advances in cell, module, and
processing technology performed primarily by
DOE/JPL-sponsored R&D.
Participation by FSA personnel having access to
the entire spectrum of PV technology and PV
field operational results including items such as
quality assurance.
The excellent cooperation established and
continued at the working level by the PV industry
and FSA personnel throughout all these technical
activities.
Figures 50 through 54 show the five block buys
and Tables 18 through 20 show their characteristics.
Establishment and continuing upgrading of design
and performance specifications based upon the
60
• FIRST LAMINATED MODULE
• CELL INTERCONNECT AND TERMINAL REDUNDANCY
• ENVIRONMENTAL TESTS LIMITED TO:
TEMPERATURE CYCLE
HUMIDITY SOAK
• QA SPECIFICATION INTRODUCED
• ELECTRICAL PERFORMANCE CRITERIA
115.8 VOLTS - 60°C CELL TEMPERATURE)
• MANY DESIGN IMPROVEMENTS
DURING PRODUCTION
• STANDARD ARRAY SIZE AND MOUNTING
• ELECTRICAL PERFORMANCE PER
MANUFACTURERS RATINGS
• INTRODUCTION OF GROUNDING SAFETY PROVISIONS
• EXPANDED ENVIRONMENTAL QUALIFICATION TESTING
• THERMAL CYCLE
• HUMIDITY CYCLE
• STRUCTURAL LOADING
Figure 51. Block 1/: 1976-1977, Designed to FSA
Specification, 127 kW
Figure 50. Block I: 1975-1976, Off-the-Shelf Design,
54 kW
• DESIGN AND TEST SPECIFICATIONS ESSENTIALLY SAME AS BLOCK II
• IMPROVEMENTS IN DESIGN AND PRODUCTION PROCESSES RESULTING
FROM BLOCK II EXPERIENCE
• MORE UNIFORM QA STANDARDS
Figure 52. Block 11/: 1978-1979, Similar
Specifications, 259 kW
61
• TYPICAL DESIGN FEATURES
• LAMINATED MODULE CONSTRUCTION
• FAULT TOLERANT CELL AND CIRCUIT DESIGNS
• LARGER POWER OUTPUT
• CELLS WITH BACK SURFACE FIELDS
• GLASS FRONT FACE
• INNOVATIVE DESIGN FEATURES
•
•
•
•
•
•
•
SHAPED CELLS
ION IMPLANTED CELLS
SEMICRYSTALLINE CELLS
ETHYLENE VINYL ACETATE ENCAPSULANT
BATTEN - SEAM ROOFING SUBSTRATE
FRAMELESS MODULE
INTEGRAL BYPASS DIODES
Figure 53. Block IV: 1980-1981: Industry Designs Reviewed by FSA, 26 kW of Prototype Modules
62
• TYPICAL DESIGN FEATURES
INNOVATIVE DESIGN FEATURES
• LARGER POWER OUTPUT
• MODULE EFFICIENCY> 10% (EXCEPT
RIBBON CELL MODULE)
• MAJOR INCREASE IN AREA AND
POWER OUTPUT
• MET MORE STRINGENT QUALIFICATION
TESTS
• GLASS TOP COVER
• ETHYLENE VINYL ACETATE
ENCAPSULANT
• LAMINATED COMPOSITE FILM BACK
COVER
• VIRTUAL ELIMINATION OF THE
FOLLOWING CATASTROPHIC FAILURE
MODES
• UNACCEPTABLE CELL CRACKS
• INTERCONNECT FAILURES
• LAMINATED MODULE CONSTRUCTION
• HOT·SPOT FAILURES
• FRAMELESS MODULE
• SHAPED CELLS (HIGHER
PACKING FACTOR)
• HAIL DAMAGE
• MODULE WITH CELLS MADE FROM
SILICON RIBBON (EFG) GROWN TO
THE CORRECT THICKNESS
• PARALLEL CELL STRINGS
FAULT TOLERANT CELL AND CIRCUIT
DESIGNS
BYPASS DIODES
Figure 54. Block V: 1981-1985: Industry Designs Reviewed by FSA, Small Module Quantities
for Evaluation Only
63
Table 18. Module Cell and Circuit Characteristics
CELL
SIZE
I
II
III
IV'
V
MANUFACTURER
SENSOR TECH.
SOLAREX
SOLAR POWER
SPECTROLAB
SENSOR TECH.
SOLAREX
SOLAR PDWER
SPECTROLAB
ARCO SOLAR
MOTOROLA
SENSOR TECH.
SOLAREX
SOLAR POWER
ARCO SOLAR
ASEC
G.E. 8
MOTOROLA
PHOTOWATT
SOLAREX
SOLAREX 8
SPIRE
ARCO SOLAR
8
G.E.
MSEC8
SOLAREX
SPIREs
MOOEl NO.
V·13·AT
7B5
E-l0·229·1.5
060513·8
20·10·1452·J
A·0221·D
E·1D008·C
022962-G
10699·C
P·0170·770-J
20·10·1646
A·0221·G
E·l0008·F
012110·E
60-3062·F
47J254977Gl·C
MSP43D40·G
ML·1961·D
560-BRC
580·BT·R·C
058·0007-A
004·014168·2
47E258449G2·A
R8·180·12-D
C·120·10A
058·0006·B
CIRCUIT
BASE
MATl
CZ
SERIES PARAllEl CROSS BY·PASS
(mm)
QNTY
JUNCTION CELLS
SHAPE
CELLS
TIES OIOOES
25
500lA ROUNO
25
NIP
18
7601A
NIP
18
22
8701A
PIN
22
500lA
20
NIP
20
44
5601A
44
42
7601A
42
40
102 OIA
PIN
40
120
500lA
NIP
40
3
7601A
41
41.
48
7601A
12
4
11
44
44
5601A
42
7601A ROUND W/1 flAT
42
40
40
102 OIA ROUND
PIN
35
103 DIA ROUND W/2 flATS
NIP
35
1
136
7601A ROUND
34
4
5
1
19
100 OIA ROUND Wll FLAT
19
NIP p+
33 100 x 100 QUASI·SQUARE
33
7601A ROUND
12
6
72
72
95 x 95 SQUARE
SEM~XTl
36
2
35
36
12
6
72
95 x 95 SQUARE
SEM~XTl
11
12
108
64 x 64 QUASI-SQUARE
CZ
36
3
11
2
72
97 x 97 QUASI·SQUARE
CZ
6
12
3
1
NIP
72 100 x 100 QUASI·SQUARE
CZ
2
34
NIP
36
3
432
95 x 48 RECTANGULAR
EFG
12
2
1
NIP
36
117 101 x 101 RECTANGULAR
SEMI·XTl NIP
9
13
1
72
91 x 91 QUASI-SQUARE
CZ
2
2
N/P·P+
36
3
t
!
t
!
NOTE: 8RESIDENTIAL MODULE
Table 19. Module Performance Characteristics
SAMPLE PERFORMANCE
AT 100 mWlem 2• AM 1.5. 28°C CEll TEMP.
MANUFACTURER
SENSOR TECH.
SOLAREX
SOLAR POWER
SPECTROLAB
SENSOR TECH.
SOLAREX
SOLAR POWER
SPECTROLAB
ARCO SOLAR
MOTOROLA
III SENSOR TECH.
SOLAREX
SOLAR POWER
ARCO SOLAR
ASEC
G.E"
IV MOTOROLA
PHOTOWATT
SOLAREX
SOLAREXa
SPIRE
ARCO SOLAR
G.E.·
a
MSEC
SOLAREX
SPIREa
MOOEL NO.
Pma ,
(WI
VP max
(VI
IP ma ,
V·13·AT
785
E·l0·229·1.5
060513·8
20·10·1452·J
A·0221·0
E·l0008·C
022962·6
10699·C
P·0170·770·J
20·10·1646
A·0221·G
E·l0008·F
012110·E
60·3062·F
47J254977GI·C
MSP43040·G
ML·1961·0
580·BHC
580·BT·R·C
058·0007·A
4.7
8.7
13.2
4.7
11.4
20.5
33.8
30.0
22.8
26.2
11.3
21.7
34.8
35.7
84.6
18.8
37.3
38.6
62.6
60.8
57.0
9.8
7.0
9.6
9.4
20.7
18.0
18.0
lB.2
18.2
5.9
20.2
17.8
18.3
16.6
16.5
8.5
16.2
5.68
16.1
5.31
16.2
0.48
1.24
1.38
0.50
0.55
1.14
1.88
1.65
1.25
4.45
0.56
1.22
1.90
. 2.15
5.11
2.21
2.30
6.79
3.90
11.4
3.52
84.1
004·014168·2
47E258449G2·A 81.7
185.
R.·180·12·0
139.
C·120·10A
70.7
058·0008·B
5.82
17.0
15.3
5.84
16.1
14.5
4.61
12.1
23.8
4.39
(AI
Ise
IAI
FILL
FACTOR
AT 100 mWiem • AM 1.5. NOCT
MOOULE
EFF. (%1
CElL
EFF.el%1
4.8
6.5
5.8
5.9
6.8
6.0
7.4
6.6
8.4
7.7
6.8
6.5
7.7
9.6
10.1
9.6
8.8
7.2
8.2
8.1
11.4
9.4
10.6
10.2
12.0
10.6
10.7
10.7
12.7
12.2
11.8
10.5
11.6
11.2
12.6
13.6
12.6
11.6
11.6
9.6
9.3
13.6
10.4
18.7
31.3
2B.5
20.6
23.6
10.2
19.7
32.2
32.4
77.4
15.3
34.3
34.9
57.3
54.5
50.6
18.7
16.3
16.6
17.3
16.5
5.3
18.6
16.4
17.2
15.0
15.0
7.1
15.1
5.10
14.4
4.70
14.2
0.56 23.4
1.15 22.4
1.89 22.0
1.65 21.9
1.25 22.0
4.45
6.6
0.55 23.0
1.20 22.1
1.87 22.0
2.16 19.6
5.16 19.2
2.16
9.6
2.27 18.4
6.84
6.5
3.98 18.1
11.6
6.2
3.58 18.6
11.3
10.5
8.4
10.3
10.1
12.6
11.7
9.4
11.7
13.3
75.0
65.4
165.
123.
62.7
5.20
13.3
13.2
5.18
14.5
14.4
4.92
12.5
23.7
4.32
24.8
0.60
0.77
24.3
0.59
1.43
23.5
0.73
1.98
23.0
0.70
1.B6
0.71
23.3
1.38
7.1
4.82
0.76
0.74
24.6
0.62
23.7
1.40
0.65
23.6
1.97
0.75
21.0
2.42
0.70
20.2
5.40
0.78
11.0
0.68
2.53
19.5
2.50
0.76
6.98 7.58
0.73
19.6
4.50
0.71
6.60 13.2
0.69
20.3
3.64
0.77
AT 100 mWlem 2• AM 1.5. '25°C CElL TEMP.
7.16 15.9
20.9
5.65
18.9 13.3
7.47 26.7
20.7
4.79
0.74
0.69
0.74
0.70
0.71
Ise
(AI
IP max
(AI
0.59
1.44
1.98
1.88
1.40
4.88
0.62
1.41
1.98
2.42
5.45
2.53
2.52
7.62
4.58
13.3
3.67
Fill
FACTOR
0.75
0.58
0.72
0.69
0.67
0.73
0.72
0.63
0.74
0.68
0.74
0.63
0.74
0.70
0.69
0.66
0.74
b
MOOULE
EFF. (%1
CEll
EFF.c(%1
6.3
5.5
6.9
6.3
7.6
7.0
6.1
5.8
7.1
8.7
9.3
7.8
8.1
6.6
7.5
7.3
10.1
9.6
9.8
9.7
11.7
11.0
10.8
9.4
10.4
10.3
11.4
12.6
10.3
10.6
10.6
8.8
8.4
11.9
39
48
49
35
43
47
46
41
50
53
43
46
46
46
47
58
49
47
49
56
49
10.1
6.4
7.5
9.1
9.0
11.2
9.3
8.4
10.3
11.8
49
65
48
46
47 d
AT 100 mWlcm AM 1.5. NOCT
6.56 16.1
17.7
5.69
17.9 13.7
6.79 27.2
18.9
4.84
0.71
0.85
0.67
0.67
0.89
NOTES: ·RESIOENTIAL MOOULE
2
bNOMINAL OPERATING CEll TEMPERATURE: CEll TEMPERATURE IN OPEN·CIRCUITEO MOOULE EXPOSEO TO 80 mW/lcm INSOLATION IN AMBIENT OF 20°C. 1 mI. WINO VELOCITY
cENCAPSULATEO CEll
dRACK·MOUNTEO
64
b
NOCT
(OCI
Table 20. Module Mechanical Characteristics
MANUFACTURER
SENSOR TECH.
i SOLAREX
SOLAR POWER
SPECTROLAB
SENSOR TECH.
if SOLAREX
SOLAR POWER
SPECTROLAB
ARCO SOLAR
MOTOROLA
Ii SENSOR TECH.
SOlAREX
SOLAR POWER
ARCO SOLAR
ASEC
G..E."
IV MOTOROLA
PHOTOWAIT
SolAREX
SOlAREX"
SPIRE
ARCO SOLAR
G.E."
MSEC"
V SOLAREX
SPIRE"
b
AR£A
Iml
0.097
V·13·AT
7B5
0.133
E·l0·229·1.5
0.229
060513·8
0.090
0.168
20·10·1452·J
A·0221·0
0.335
E·l0008·C
0.454
022962·G
0.453
10699·C
0.270
0.340
P·0170·770·J
20·10·1646
0.166
A·0221·G
0.335
E·l0008·F
0.454
012110.£
0.372
SO·308H
0.834
47J254977G~C
0.196
MSP43040-o
0.426
ML·1961·0
0.532
58D-BH·C
0.782
0.749
580·BT·RoC
058·0007·A
0.504
0.745
lr04=014168·2
47E256449G2·A 0.778
2.154
R..180·12·0
1.331
C·120·10A
0.675
058.flO08·6
MODel NO.
"RESIDENTIAL MODULE
bexPOSEO AREA
COVERALL DIMENSION
LENGTH WIDTH
t
(m!t
Iml
0.57
0.51
0.61
0.66
0.582
0.579
1.168
1.168
1.168
0.583
0.582
0.579
1.168
1.219
1.199
0.618
1.198
1.199
1.200
1.193
1.200
TIff
1.228
1.791
1.391
1.134
0.17
0.26
0.37
0.12
0.289
0.579
0.389
0.388
0.231
0.583
0.286
0.579
0.389
0.305
0.696
0.669
0.356
0.444
0.635
0.628
0.417
0.610
0.633
1.203
0.957
0.595
MASS
(kg!
1.3
1.1
2.6
1.6
1.5
4.1
7.6
6.1
3.7
6.6
3.7
4.4
7.4
5.2
13.5
4.0
5.8
7.4
13.9
11.2
7.8
12.0
13.6
29.5
23.6
7.3
SUPERSTRATE
OR TOP COVER
RTY-615
SYlGARO 184
D.C. R4·3117
GLASS
RTV-S15
SYLGARO 184
D.C. XL·2577
GLASS
Rrt615
SYLGARO 184
D.C. R4·3117
GLASS
SUBSTRATE
OR BOTTOM COVER
ALUMINUM
NEMA·G 10 BOARD
NEMA·G 10 BOARD
ALUMINUM
ALUMINUM
NEMA·G 10 80ARO
GfR POLYESTER BOARD
MYlAR
TEOLAR
STAINLESS STEEL
ALUMINUM
NEMA·G 10 BOARD
GFR POLYESTER BOARD
TED/SIlTED
TEOLAR
MEAD PANHOARO
TEOIALITEO
TEO/AlITEO
TEOLAR
TEOLAR
MYLAR·AL·COAT
TEO/PEIITEIi"
TEOIPET/ALITEOd••
PET/ALITEO·
PET/MYLAR/TEO·
TEOLAR
"PLus SHINGLE MATERIAL
"PET. POlYESTER FILM. POlYTHYLENE TEREPHTHALATE
65
ENCAPSULANT
RTV·615
SYLGARO 184
SYlGARO 184
RTY-615
RTV·615
SYLGARO 184
SYLGARO 184
PV8
PVB
0.C.03·6527A
RTV·615
SYLGARO 184
SYlGARO 184
PV8
PV8
G.E. SCS2402
PV6
PV8
EVA
ENCAPSULANT
METHOD
FRAME
CASTING
NONE
LAMINATION
LAMINATiON
G
CAsr
LAMINATION
ALUM.
NONE
ALUM.
AlUM.
ST. STEEl
NONE
ALUM.
NONE
AlUM.
ALUM..
NONE
ST. STEEL
ALUM.
ALUM.
NONE
ST. STEEl
ALUM.
NONE
!
ElECTRICAL
CONNECTIONS
PACKING
FACTOR
TERMINALS
PIGTAILS
J·BOX/CABLE
TERMINALS
TERMINALS
J.BOX
J·BOX
PlUG·IN
TERMINALS
0.51
0.61
0.57
0.49
0.64
0.56
0.69
0.52
0.69
0.65
0.65
0.56
0.69
0.76
0.74
0.76
0.76
0.62
0.85
0.87
0.85
0.90
0.90
0.89
0.88
0.76
t
J·BOX
I
PllTAILS
FlAT·CABLE
J·60X
PlUG·IN
PIGTAILS
PIGTAILS
PlUG·IN
J·BOX
FLAT CABLE
J·BOX
PLUG·IN
PLUG·IN
systems. Their size was too small, the mechanical
interfaces (bolt patterns) varied from module to
module (even within the same manufacturers model,
they were unsafe electrically for use at high
voltages), and they lacked means of electrically
interconnecting them cost effectively" To define the
needed changes, the FSA Project worked closely
with the systems and applications researchers at
NASA Lewis Research Center and Sandia National
Laboratory to understand the problems being
uncovered and the future applications being
identified. Next, FSA researchers contracted with
leading industrial teams, knowledgable of the detailed
technical characteristics of future applications, to
identify and develop concepts and requirements for
flat-plate PV modules specifically suited to these
uses. Large architect and engineering (A&E) firms
with extensive experience building large central utility
power plants defined optimum modules and arrays
for utility-scale applications; simultaneously,
residential and commercial architectural firms
developed the detailed requirements and concepts
appropriate for these sectors, including consideration
of applicable building and safety codes and labor
practices.
Engineering Sciences
and Design
At the start of the FSA Project, there was a need
to establish a complete new PV module and array
technology for terrestrial applications. The use of
solar cells in space was a very different application
than any found on Earth. Thus, there was a need to
define potential terrestrial applications, and to
develop the operational and environmental
requirements and technology needed for the
manufacture and use of PV products in these
applications.
The objective is to develop reliable, low-cost
module and array engineering and design
technologies for integrating cells and modules into
cost-effective and safe arrays for large-scale
applications.
• Identify and develop functional requirements
and specifications to ensure module
compatibilty with the various operational and
safety environments of future PV applications.
tl
As the early deficiencies were gradually resolved,
the above activities coalesced into a coherent array
subsystem design effort (Figure 55) addressed to
optimizing the overall module and array designs for
the intended applications so as to reduce costs and
increase efficiency and lifetime. The environmental
conditions that PV modules "and arrays are subjected
to during field operation were assessed, measured,
and defined. This covered mean operating conditions
in addition to extremes for conditions such as UV/
solar weathering, temperature, temperature cycling,
hail impact, and combinations of the above (Figure 56). Means of quantifying and reducing the
operating temperatures of modules were developed;
detailed wind tunnel analyses were conducted to
define expected wind loading levels, the effectiveness
of wind barriers, and dynamic flutter magnitudes.
Based on the structural loading studies, new module
structural designs were developed. These designs
included frameless modules and low-cost groundmounted array structures (Figure 57). A new
methodology for determining the breaking strength
of glass was also developed.
Develop engineering design methods and data
and advanced safe, reliable design concepts
focused at the needs of future large-scale PV
applications.
• Verify the adaptability of the technology to meet
required operational conditions.
Background
The use of crystalline silicon solar cells in space
had demonstrated that PV power systems were a
practical and reliable method of generating electrical
power. The solar cells and the "solar panels," as
designed, were light-weight, resistant to radiation,
capable of withstanding rapid heating and cooling,
and were cost effective as used in space. The idea
of using photovoltaics as a major power source in
the terrestrial environment meant that the new
functional requirements of these applications had to
be identified and the necessary technologies
developed. Not only were the environmental
requirements significantly different with the addition
of wind and moisture-related weathering, but the
institutional requirements were vastly changed with
added considerations of building and safety codes,
user interaction, product liability concerns, and
interaction with the community of installation trades
and operation and maintenance personnel.
To improve the reliability of the arrays, complex
circuit analyses were used to define optimum seriesparallel electrical circuit redundancy approaches at
the array and module level, and optimum maintenance and replacement trade-offs (Figure 58); other
design studies explored low-eost electrical interconnection techniques.
Project Activities: 1975 to 1981
One of the key institutional barriers uncovered in
the electrical circuit studies was the unique electrical
safety attributes of PV which are inconsistent with
many standard safety practices. Unlike conventional
electrical sources, PV modules cannot be turned off
and they cannot generate the overcurrent needed to
Deficiencies and failures were prevalent in the first
(pre-1975) terrestrial modules. In addition to the early
reliability problems, the modules were found to be ill
suited for use in the first kilowatt-scale, high-voltage
66
1975 TO 1984
Year
Technology
75
76
77
Thermal Design (Groundl
Thermal Design (Rooll
PV-Thermal Modules
Gnd-Mtd Array Structures
Residential Arrays
Series-Parallel Analysis
Bypass Diode Integration
c: "::::J1
Electrical Connectors
Module Safety Design
Array Safety Systems
Flammability
PC Interface
c
C
C
70
79
00
01
~
c::
OJ
02
c ::3.
c::: t:::::l
=
1=::::;&.
c
c f------ P
~
c
04
c==:>ll
f= :iii.
j:::;a:
= fa
0\.
= :a.
tl=*:
.., = Major Contribution
IA
=
Significant Contribution
Figure 55. Engineering Sciences Developments
HAIL IMPACT RESEARCH
OBJECTIVES
•
DEVELDPTESTREOUIREMENTLEVELS
AND DURATlDNS,TEST PROCOURES
AND APPARATUS FOR DETECTING
~~~~f~~':,.,~ATION MODES CR'
•
~~~~~~~:~J~st!o~tG';:~~:E~~gH
g~u~~g~~F::llc~i:~~~CHANISMS
~~
• DEVELOP STATISTICAL DATA DefiNING
PROBABIlITV OF IMPACT BV VARIOUS
HAIL SIZES IN GEOGRAPHIC REGIONS OF
UNITEOSTATES
• DETERMINE RELATIVE TOLERANCE OF
MODULE DESIGNS TO SPECIFIEO ENVI
RONMENTAL STRESSES AND IDENTIFV
~~~~~~~t:AMETERS INFLUENCING
CLEAR SILICONE POTTING
o 10_in
• CONDUCT MODULE HAIL RESISTANCE
CDSTI8ENEfITASSESSMENT
PHILOSOPHY
012·;n ANNEALED GLASS
• CELLS IESPEClALl V EDGES NEAR ELEC
TRlCALCONTACTSI
MINIMUM CHANGE IN TESTS TO PERMIT
LONG TERM CORRELATIONS
0.12-ln. TEMPERED GLASS
• ENCAPSULANT SYSTEM ICORNERS ANO
EOGES, POINTS OFSUPERSTRATE
~~~~~RF~6~I~J~E~~~R~~I~SUU~~~1TI
• MINIMUM TEST COMPLEXITY AND DURA
TlON TO REDUCE COST
• EASILV IMPLEMENTED flV MANUFACTUR
~:~J~~~ COMMERCiAl TESTING ORGANI
FAILURE RANGE
NO DAMAGE
ACRVlIC SHEET
009_in.ANNEALEOGLASS~.
lALUMSUBSTRATEI
SUSCEPTIBLE PARTS
• MAXIMUM TEST REPEATABILITY TO
ALLOW CORRELATION AND COMPARISON
•
HAil IMPACT RESISTANCE
OBJECTIVES
HAIL IMPACT DAMAGE
ON DEVELOPMENTAL MODULE
!3/4-in.ICE8ALLI
01B_in. TEMPERED GLASS
o
O.S
1.0
1.5
2.0
IfAILSTONEOIAMETER lin I
EXPERIMENTAL HAIL RESISTANCE
TEST APPARATUS
Figure 56. Environmental Requirements Example
BURIED TRUSS STRUCTURE
4' X 4' MODULE
OBJECTIVE
DEFINITION OF ARRAY DESIGN
TRADE-OFFS AND GENERATION
OF DESIGN GUIDELINES, REQUIREMENTS, AND CONCEPTS
APPROACH
• DEVELOP ADVANCED
DESIGN CRITERIA
• DEVELOP ANALYSIS TOOLS
• TESTS
• EQUATIONS
• COMPUTER PROGRAMS
• GENERATE REPRESENTATIVE
DESIGNS AND DESIGN
TRADE-OFF DATA
TRUSS STRUCTURE IS BURIED IN
1 1/2' WIDE x 31/2' DEEP x 9'LONG
TRENCHES
REFILL AND TAMP TRENCHES USING
REMOVED EARTH
PROTOTYPE STRUCTURE
VIBRATION TEST
Figure 57. Advanced Array Structures
blow fuses or circuit breakers in the event of a short
circuit or short to ground. To develop the needed
safety technology, JPL contracted with Underwriters
Laboratorieswho developed detailed module safety
requirements and conceptual approaches to the
entire electrical safety system for a complete PV
power system (Figure 59). On the basis of this work,
requirements for UL listing of modules have been
developed, and a new Article 690, covering required
electrical safety features in high-voltage PV power
systems, has been included in the 1984 revision of
the National Electrical Code.
67
RELIABILITY/DURABILITY
OBJECTIVE
• DEVELOP RELIABILITY OATA, DESIGN GUIOElINES. AND REQUIREMENTS FOR
flAT ARRAYS AND MODULES
I
SERIES/PARALLEL NOMENCLATURE
GENERALIZED REtlABltlT'1'/DURAllILlTY GOAt
liFE-CYCLE ECONOMIC PERFORMANCE
SHALLDE EQUIVALENT TO $O.101W
100EFFlCIENCY. NO OEGRADATION FOil 20 VUfO
APPROACH
• DEVELOP SERIES PARALLEL ElECTRICAL RELIABILITY MODEL
• DETERMINE ARRAY SENSITIVITY TO KNOWN COMPONENT FAILURE MOOES
• DEVELOP ARRAY/MODULE CIRCUIT CONFIGURATION GUIDELINES TO MINIMIZE
ARRAY DEGRADATION DUE TO COMPONENT FAILURES
• GENERATE TEST TECHNIQUES AND ACQUIRE RELIABILITY DATA FOR MOOULE
COMPONENTS. (CEllS SOLDER JOINTS. CONNECTORS, .llcJ
• OEFINE COST _ BENEFITS ASSOCIAUO WITH COMPONENT RELIABILITY
IMPIlOVEMENTS
• DEFINE MAINTENANCE PRACTICES FOR MINIMUM LIFE
CYCLE COSTS
BASELINE
NORMAU2EO[
POWER
OUTPUT
ARRAY POWER DEGRADATION vs
CIRCUIT REDUNDANCY
r->
3 PARALLEL STfllNGS
MOOU"
2 SERIES BlElCKS
2 CEllS PER SUBSTRING
2 DIODES PER MO.DULE---~-
[LaUIVALENT
,;--~",-~-j20
.
BRANCH CIRCUIT
3 PARALLEl STRINGS
6 SERIES BtOCKS
2 CELLS PER SUBSTRING
1 DIODE PER SERIES llLOCK
30
_
L1FE·CYCLE ENERGY COST vs CIRCUIT CONFIGURATION
CEll FAILURE RATE ~ 00001 PER YEAR
8 PARALLEL X 2400 SERIES X N SERIES BLOCKS
"~
"o".,,"s.,oc..
"
0.5
.
,
o
5
10
T1ME!YEARSI
15
20
Figure 58. Electrical Circuit Reliability
FRAME/--=STRUCTURE
GROUNO
applications. Detailed studies were conducted
characterizing the operating temperatures of roofmounted modules and defining residential array
design guidelines including array wiring practices and
bypass-diode packaging concepts. Testing of the new
environmentally durable encapsulants indicated
substantial progress in the area of reliability, but a
step backward relative to flammability and suitability
for module applications with a high fire risk.
Research was initiated to characterize the flame
resistance of new module designs and to develop
flame-resistant design approaches and encapsulant
materials (Figure 60).
CIRCUIT GROUNO
Figure 59. Isolation/Grounding Safety Concepts
The most important technique for getting modulemanufacturing industry participation in the
modulelarray design process and for transferring new
knowledge into practice, was to make periodic
procurements of modules. These purchases of
advanced-design modules to successively more
definitive and more appropriate requirements, and
the subjection of these modules to qualification life
tests and field tests, resulted in a very dynamic and
fruitful cooperative industry/Government effort.
There were a number of design specifications that
evolved over the years, each building on the previous
one together with the technology developments and
the application and testing experience. The latest
Block V Module Design and Test Specification was
issued in two versions in 1981; one specification is
for Residential Applications, and the second is for
Intermediate Load Applications, i.e., general purpose
uses.
FLAMMABILITY RESEARCH
Figure 60. Flammability Research on
Rooftop Module
Project Activities: 1981 to Present
Since 1981, the evolution of module and array
requirements and technology has continued with
emphasis on completing the technology base for
residential applications and supporting the design
and fielding of the first megawatt-scale utility
The need for improved integration between the
array and the power conditioner led to extensive
modeling of array output versus time for various sites
in the United States. Characterizations included
68
statistical data on maximum array voltage, current
and power levels and the performance efficiency of
various maximum-power tracking algorithms.
With the gradual completion of the engineering
technologies for crystalline-silicon modules,
Engineering Sciences research has gradually phased
down in recent years and turned to the needs of the
emerging new thin-film modules. These needs seem
to be well served by the technology base resulting
from the 10 years of research on crystalline-silicon
modules and arrays.
It
Module electrical circuit design guildelines
including interconnect redundancy and cell
contacting techniques
It
Module electrical terminal needs and designs
@
It
Module electrical insulation system design
guidelines and testing techniques
Module and array thermal design guidelines for
cooler operation, resulting in increased power
output and longer life
Significant Accomplishments
@
Specific design data and recommendations that
have evolved include:
It
@
Module design specifications including
environmental endurance test requirements
@
@
It
@
It
@
It
Module electrical safety design requirements
and practices (U L 1703)
Standardized module thermal testing methods
Fire resistant module designs and encapsulant
materials
Residential array design approaches
It
Low-cost ground-mounted array design
approaches including frameless modules
ED
Guidelines for optimum maintenance/
replacement of failed modules in the field
ED
Data on array cleaning costs and automated
washing techniques
It
Wind pressure loads on flat modules and arrays
Array wire selection and safety design
guidelines
ED
Structural capabilities of glass as used in
modules, and glass sizing algorithms
Array series/parallel electrical circuit design
guidelines including grounding and bypass
diode design guidelines
ED
Design guidelines for optimally interfacing
arrays with power conditioners
Safety system design concepts and
recommendations (National Electrical Code
Article 690)
Detailed assessments of residential and
commercial bUilding codes and the implications
on the use of photovoltaics
Module product-liability guidelines
69
Encapsulation
assurance, for example, requirements for and control
of encapsulant curing. The product quantities were
so small that manufacturers could not afford to do a
thorough job of studying materials and processes. A
comprehensive long-range encapsulation technology
plan was needed and was formulated by the FSA
Project.
In 1975, the encapsulation requirements of PV cells
for terrestrial use had not been comprehensively
assessed, needs had not been defined, nor potential
low-cost encapsulation materials investigated.
Therefore, the FSA Project set out to assess the
needs in order to identify, and/or develop new
materials and new material technologies that were
inexpensive and could protect solar cells for years.
The assessment led to the establishment of
encapsulation needs, goals, and detail requirements
and a plan on how to meet the goals.
Project Activities: 1975 to 1981
In the implementation of the plan to meet the
objectives and goals, five activities were followed: (1)
the generation of specifications and functional
requirements for encapsulation materials; (2) the
identification or development of low-cost materials
satisfying the specifications and functional
requirements; (3) engineering material usage; (4)
recognition of life and/or weathering deficiencies in
the low-cost materials; (5) generation of necessary
design approaches or material modifications to
enhance life or weathering stability; and (6) lifeprediction methodologies for encapsulation systems.
The objectives of the encapsulation activities were
to define, develop, demonstrate, and qualify
encapsulation systems, and materials and processes
that meet the FSA Project module life, cost, and
performance goals; and to develop and validate
encapsulation system life prediction: methodology
based on modeling life-limiting failure modes and on
conducting and analyzing aging tests.
Two basic encapsulation systems for terrestrial
modules evolved: substrate and superstrate
(Figure 61). This nomenclature refers to the method
of structurally supporting the encapsulated cells and
their electrical interconnections, i.e., sub or bottom
side and super or top sunnyside. Today, the superstrate design is universally used because tempered
glass is a practical top cover for use with crystalline
silicon solar cells. Thin-film cells have the potential
for use in flexible modules; however, a practical
flexible top cover is unavailable. Consequently, the
preponderance of the effort, to date, has centered
on the technology for laminating the active components in a PV module and their function are shown
in Figure 62.
The goals are to develop encapsulation system
technology adequate for:
• Module life of 30 years (increased from 20
years).
• Encapsulation materials, including substrate
and/or superstrate, costs not exceeding $14/m 2
(1980 dollars).
• Initial optical transmission of 90% and a loss of
less than 5 % after 20 years of use.
• Capability to withstand an electrical breakdown
voltage of 3000 V dc.
The key component is the pottant, which is a
transparent, elastomeric material that encases the
cells and their electrical circuitry. In the fabricated
module, the pottant provides:
• Structural integrity and durability to withstand
handling and weather.
• Cost-effective processing in an automated
factory with high yields.
• Maximum optical transmission in the silicon
solar cell operating wavelength range of 0.4 to
1.1 f-Lm.
Background
Typical terrestrial PV module fabrication prior to
1975 was by hand assembly. The solar cells were
interconnected by hand soldering. The strings of cells
were encased by hand-poured silicone rubber, which
usually resulted in non-uniform products. The
durability of a module was unpredictable, often with
delamination and/or cracking of the encapsulant. The
module materials were expensive and the
encapsulation concepts, designs, and processes
were not amenable to mass production. Consistent
module quality was a problem because of a lack of
good processing knowledge, capabilities, and quality
70
Ii
Retention of a required level of electrical
insulation to protect against electrical
breakdown, arcing, etc., with the associated
dangers and hazards of electrical fires, and
human safety.
Ii
The mechanical properties to maintain spatial
containment of the solar cells and
interconnects, and to resist mechanical creep.
The level of mechanical properties also must
not exceed values that would impose undue
mechanical stresses on the solar cells.
DES!GN
COST EST
(~J",~L
GLASS (STRUCTURAL)
$5.00-10.00
SPACERiSCRIM
0.100.10
PODANT
0.70-2.00
SMeERISCRIM
0.10-0.20
", :~:ifj~Ci~SOLAR CELLS
PODANT
BACK COVER fiLM
SUBTOTAL
COST EST
1_2 ><1.2 m SIZE
$1.10-2JlOlm2
Figure 61. Encapsulation System Designs
must also be resistant to atmospheric soiling, and be
abrasion resistant, anti-reflective, and easily
cleanable.
Also, the pottant must be commercially available
and be readily processible for automated fabrication.
Because there is a significant difference between the
thermal-expansion coefficients of polymeric materials
and the silicon cells and metallic interconnects,
stresses developed from the thousands of daily
thermal cycles can result in fractured cells, broken
interconnects, or cracks and separations in the
pottant material. To avoid these problems, the pottant
material must not overstress the cell and
interconnects, and must itself be resistant to
fracture. From the results of a theoretical analysis,
experimental efforts, and studies of available
materials, the pottant must be a low-modulus
elastomeric material.
The edge of the laminate needs to be protected so
that delamination does not begin and destroy the
module. A gasket is used if the module is to have a
frame or it may be used for the installation of
frameless modules.
The top cover is in direct contact with all of the
weathering elements: ultraviolet, humidity, water,
oxygen, sunlight, etc., and must protect underlying
materials from degrading effects. The outer surface
During module fabrication, primers are required to
obtain good adhesion of the polymeric pottant to
glass, solar cells, and polymeric films that are used
for back covers.
The back cover protects the cells, interconnects,
pottant, and the substrate, if there is one, from
weather and from electrical shorting. The cover
should be weatherable, hard, mechanically durable
and tough, and have a white, high-emissiVity surface
for cool module operation.
71
MODULE ENCAPSULATION
FUNCTIONAL DESIGN ELEMENTS
MODULE ELEMENTS
REQUIRED FUNCTIONS/CHARACATERISTICS
TOP SURFACE
(AS IS. MODIFIED.
OR COATED)
_ T O P COVER
-
•
•
•
•
----
SPACER, POROUS
{
• MAXIMUM LIGHT TRANSMISSION
• UV SCREEN
• STRUCTURAL SUPERSTRATE
(IF REQUIRED)
- - - f . AIR RELEASE
l . MECHANICAL SEPARATION
-POTTANT-----~
-
LOW SOILING
EASY CLEANABILITY
ABRASION RESISTANT
ANTI REFLECTIVE
• ENCASES SOLAR CELLS
AND INTERCONNECTIONS
• HIGH LIGHT TRANSMISSION
SOLAR CELLS
-PODANT
-
SPACER, POROUS
-
DIELEC:rRIC - - - - - ( • ELECTRICAL ISOLATION
-
SUBSTRATE - - - - - {
-
BACK C O V E R - - - - { • BACKSIDE PHYSICAL PROTECTION
BACKSIDE WEATHERING BARRIER
PRIMERS/ADHESIVES
(AS REQUIRED)
• STRUCTURAL SUPPORT
(IF REQUIRED)
l .
COMPLETED MODULE -
{
• GOOD BONDING
• GOOD HEAT TRANSFER
• GOOD PROCESSABILITY
Figure 62. Module Encapsulation Functional Design Elements
temperatures up to 80°C), except for ultraviolet
weathering. Of the four intermediate cost polymers,
two were casting liquids and two were dry films that
could be used for fabricating laminated modules.
The dry films are practical for PV module uses
because of the ease and simplicity of handling and
processing. Initially, ethylene vinyl acetate (EVA) was
the only low-cost dry film available; therefore,
emphasis was placed upon efforts to adopt it for PV
use. The search for a second dry film polymer
continued in case EVA did not fulfill the encapsulation
requirements. Ethylene methyl acrylate (EMA) was
discovered 2 or 3 years later, but by then it seemed
that EVA would satisfy the needs. A lack of sufficient
funds inhibited a thorough investigation of EMA
although, to date, there are no known EMA
deficiencies for use as a PV module pottant. The
approach used was to develop an EVA that could be
thoroughly evaluated while work progressed on an
advanced EVA.
Many of the low-cost polymeric materials that
satisfy module engineering and encapsulation
processing requirements are not intrinsically weather
stable. Therefore, emphasis has been applied to
specific efforts to overcome life andlor weathering
deficiencies, to enhance weather stability, and to
develop accelerated aging techniques. Specific items
include the development of chemically attachable
stabilization additives (ultraviolet screening agents,
anti-oxidants, etc.), computer-assisted kinetic
modeling of outdoor weathering reactions, and the
use of new and novel outdoor-heating racks and
controlled-environment reactors as accelerated aging
techniques.
Transparent polymeric pottants are subject to
three weathering actions: ultraviolet reactions,
thermal oxidation, and hydrolysis. During the
assessment of various pottants, each polymer was
studied for its ability to withstand the above
weathering actions and the cost of the product. It
was determined that all the lowest cost polymers
were subject to all three weathering actions, a
number of expensive polymers were subject to no
weathering actions, and a few intermediate cost
polymers were resistant to weathering (at
Other encapsulation topics addressed related to
life and durability included: (1) soiling, (2) electrical
insulation, (3) primers and adhesives, (4) encapsulant
system design analysis, and (5) material
characterization (see Figures 63 and 64).
72
ENCAPSULANT SYSTEM DESIGN ANALYSIS
RESPONSE OF MODULE COMPONENTS TO ENVIRONMENT
O~
WILL PROVIDE CRITERIA
FOR MATERIAL SELECTION
AND MODULE DESIGN
~~
WIND: 100 mph
TEMPERATURE: _40° + 90°C
HUMIDITY: 0° - 100%
SOLAR: 100 mW/cm2
~
-
SOLAR
CELL
STRESS
POTTANT THICKNESS REQUIRED FOR SPECIFIC MODULE DESIGN AND ENVIRONMENT
Figure 63. Encapsulant System Design Analysis
MATERIAL CHARACTERIZATION
COVER SCREEN ABSORPTION OF HARMFUL UV RADIATION:
REQUIRED FOR MATERIALS SELECTION AND LIFE ASSESSMENT
1001---_
80
SOLAR
ABSORPTION
(%)
60
UV SCREEN
GUT-OFF
CURVES
(VINYl TINUVIN IN MMAj
IKORADj
40
20
o L..
)•. _~._~_~.~-4a.
300
350
400
- J_ _
_..J
450
WAVELENGTH, A(nm)
Figure 64. Material Characterization
Project Activities: 1981 to Present
verify the life of the pottant and of a complete
encapsulation system using this pottant.
The PV industry has been using a first-generation
EVA in commercial modules for more than 5 years.
This material has proven to be a surprisingly stable
polymer when correct module design and processing
have been used. The life-limiting characteristics of
this EVA have been found from experiments,
qualification testing, field operation, and accelerated
testing. With this knowledge, an advanced EVA that
corrects minor deficiencies of the earlier EVA has
been devised. On the basis of the results of 1 year of
evaluating the advanced EVA, it seems to have
corrected the previous materials deficiencies. At this
time, there are no known reasons why this new EVA
should not satisfy the requirements for a 30-year
module life. However, testing needs to be done to
In experimental studies of EVA aging, it seems
that EVA does not undergo thermal oxidation at
temperatures up to 50 DC. However, it does undergo
photo-oxidation. EVA, which yellows and degrades
within 1000 h when subjected to normal weather
conditions at 50 DC, did not yellow or degrade in
20,000 h when protected by an ultraviolet screening
film. Therefore, the life of an EVA pottant in outdoor
service at 50 DC (or less) is directly related to the
permanence of the module's protection from
ultraviolet.
The advanced EVA material (formulated by
Springborn Laboratories) uses a curing agent (TBEC)
73
ADVANCED ADDITIVES DEVELOPMENT
IMPROVED ADVANCED EVA FORMULATION
MATERIALS IMPROVEMENT USING ADDITIVES
COMPONENT
• EVA -
DEVELOPMENT OF CHEMICALLY BOUND
UV SCREENING ADDITIVES
HCH3H H
COPOLYM...:.ER_IZE_ _
WITH METHYL
METHACRYLATE (MMA)
DUPONT ELVAX 150
• PEROXIDE CURING AGENT
REPLACING LUPERSOL 101
• FASTER CURE AT LOWER
TEMPERATURES
• IMPROVED LONG·TERM
STORAGE STABILITY
• UV·2098 (AM. CYANAMID)
• UV SCREENING CHEMICALLY ATTACHABLE
FOR BETTER WEATHERING
PERMANENCE
• UV·3346 (AM. CYANAMID)
• UV ANTI·OXIDANT, A HINDERED
AMINE LIGHT STABILIZER
(HALS) HIGH MOLECULAR
WEIGHT ADDITIVE FOR
WEATHERING
PERMANENCE
H/
/~,?/~,~/~ ~~
~
COaCH, @:(:N ....
CHEMICALLY ATTACHED UV SCREEN
IN ACRYLIC (MMA) CHAIN FOR
USE IN TRANSPARENT COVER FILM
OR POTlANT MATERIALS
CONCEPT DEVELOPMENTS IN
ADVANCED STABILIZATION
ADDITIVES PROVIDED DIRECTIONS
FOR A MORE WEATHER STABLE,
EASIER TO PROCESS, EVA FORMULATION
LED TO
• BASIC RESIN
• LUPERSOL TBEC (PENN WALT)
OH
VINYL TINUVIN
(SYNTHESIZED AT
UNIV. OF MASS.)
FUNCTION
:>
Figure 65. Advanced Encapsulation Materials Research
which cures faster at a lower temperature in addition
to having a longer storage or shelf life; it also has a
chemically attachable, nondepleting ultraviolet screen
component and has a nondepleting polymeric weather
stabilizer (see Figure 65). The test results to date
have been excellent.
A definition of the intrinsic dielectric strength of
insulating materials which can be considered as a
fundamental material property has been proposed.
This is being used to quantify the aging characteristics of pottants in regard to their electrical insulation
properties.
The components of an encapsulation system need
to be bonded to each other permanently. Significant
advances have been made in this chemical bonding
technology during the past 10 years and especially
during the last 2 years when the number of adhesives
recommended for use in the PV industry has been
reduced from 23 to 3 and the adhesives themselves
are of improved quality (see Figure 66).
Predictive aging models that connect change of
encapsulation material properties and rates of
changes to module performance and life have been
developed. A correlation of accelerated outdoor aging
to normal outdoor aging by elevating the temperature
of encapsulants has been determined (see Figure 67).
Equipment for accelerated testing has also been
developed.
PROBLEM
SOLUTION
DELAMINATION OF
MODULE INTERFACES
DUE TO OUTDOOR
WEATHERING
GENERATE ACTUAL CHEMICAL BONDS
BETWEEN MATERIAL INTERFACES;
SUCH AS, EVA AND GLASS, BY EMPLOYING
SILANE CHEMISTRY, WHICH RESULTS IN
HIGH BOND STRENGTH, AND LONG·TERM
OUTDOOR WEATHER STABILITY
Power loss of modules caused by 3 years of outdoor soiling has been reduced significantly by the use
of fluorocarbon coatings on the module top surface.
On glass surfaces, this amounted to a time-averaged
loss reduction of from 2.65 to 1.55%.
The use of flexible top cover materials is receiving
increased interest by thin-film module manufacturers .
However, there is no proven top layer material other
than glass. It is the judgment of Project personnel that
neither Tedlar nor Acrylar materials, as envisioned,
will become an acceptable top material. Present ultraviolet additives to Tedlar are depleted too quickly and
the mechanical properties of Acrylar are deficient for
handling, processing, and in-field operation. Years
ago, the study of substrate encapsulation systems
was downgraded so that, with the available funds,
superstrate encapsulation could be emphasized. If
flexible module designs are to become a reality, a
new film material must be developed for use as a top
layer, combining properties such as long-term life,
low-cost, weatherability, ultraviolet permanence, good
transparency in the visible spectrum, and good
characteristics during film processing, handling, and
encapsulation processing.
o SILANE
,
•
CHEMICALLY BONDED
INTERFACE
I
I•
EVA
INTERPHASE
,I
0 - SIO-SIO-SIO-SI-
BULK
EVA
O-SIO-SIO-SI-
MOLECULAR·LEVEL MODEL
OF A SILANE (CHEMICAL)
BONDED INTERFACE GLASS AND EVA
Figure 66. Chemical Bonding Research
74
INCREASING THE SUNS INTENSITY
(8 SUNS), AT DSET-ARIZONA
INCREASING TEST TEMPERATURE AT 1 SUN,
STRINGBORN LABS, ENFIELD, CONNECTICUT
• UNITS ARE CALLED
OUTDOOR PHOTOTHERMAL
AGING REACTORS (OPTAR)
• UNITS OPERATE AT
70,90, AND 105°C
• IN A TRIAL RUN, UNITS
SUCCESSFULLY YIELDED
ACCELERATING AGING
RATES OF POLYPROPYLENE
AT 70, 90, AND 105°C,
WHICH EXTRAPOLATED
TO THE ACTUAL AGING
TIME OF POLYPROPYLENE
AT ROOM TEMPERATURES
THIS INITIAL FSA LEARNING
EXPERIENCE EVOLVED TO A
2ND GENERATION OUTDOOR
AGING CONCEPT
NATURAL OUTDOOR AGING
OF POLYPROPYLENE
(J)
(Jw
W
.a:
wo
a:w~
:::lew
!CC(J)~
a:<Ca:
wzo
D.;: ::Eo !Z
w::c
w
I(J) (J
o .
" .
110
°
l!r\..
90
70
}
ACCELERATED
TESTS
•
'\.
I"
'" ......
40
z~
-~
o~
<CI-
:::.
20
"'
<!>o
200
500
2000
0-
I-- ACTUAL
AGING TIME, HOURS
Figure 67. Outdoor Accelerated Aging Testing of Module Designs and Low-Cost Encapsulation Materials
Table 21. Encapsulation Materials-1985
PROJECTED COST IN
PRODUCTION QUANTITIES
$/m 2 (1980 DOLLARS)
THICKNESS (mm)*
LOW·IRON TEMPERED (SUNADEX)
STRONGEST, BUT DEFECT SENSITIVE
5.50· 8.50
LOW.IAQN ANNEALED
THICKNESS DETERMINED BY HAIL OR WIND
5.00· 6.00
3.0·5.0
3.0·5.0
DuPONT UV SCREENING FILM
3M UV SCREENING FILM
1.00 - 3.00
0.50· 0.70
0.025· 0.10
0.05 ·0.08
109
1
1
MODULE ELEMENT
MATERIAL
STATUS
NOTES
SUPEASTRATE
(STRUCTURAL)
GLASS -
PLASTIC FILMS
(UV SCREEN)
FLUOROCARBON - (TEDLAA)
ACRYLIC - (ACRYLAR)
PLASTIC FILMS
(NON·UV SCREEN)
FLUOREX
PAlMERS
A-11861
..
FOR BONDING EVA TO GLASS
ANTI·REFLECTION TREATMENT FOR GLASS
...
BEING DEVELOPED
SURFACE TREATMENTS
-
..
..
...
FEP
"
ANTI·SOILING TREATMENTS
..
ETHYLENE VINYL ACETATE (EVA)
ETHYLENE METHYL ACRYLATE (EMA)
EXPERIMENTAL FLUOROCARBON FILM
0
HIGH COST CASTING lI0UID
SILICONE RUBBER (RTV 615. SYLGARD)
POL YVINYL BUTERAL (PVB SAFLEX)
POTTANT
EXCELLENT WEATHER RESISTANCE
..
THERMOPLASTIC IN SAFETY GLASS LAMINATION
MELTS AND CURES DURING LAMINATION
SOFTENS AND CURES DURING LAMINATION
®
ALIPHATIC POL YETHER URETHANE
SPACER
NON-WOVEN GLASS MAT (CRANEGLAS)
FOR AIR RELEASE AND ELECTRICAL ISOLATION SPACING
SUBSTRATE
(STRUCTURAL)
GLASS. TEMPERED OR ANNEALED
STEEL
ALUMINUM (SHEET AND EXTRUSIONS)
DIELECTRIC AND THERMAL MATCH
BACK COVER
POLYESTER FILM
..
(MYLAR)
FLUOROCARBON, PIGMENTED (TEDLAR)
ALUMINUM FOIUPOLYMER LAMINATES
STEEL FOIUPOLYMER LAMINATES
FIBER GLASS COMPOSITE
..
..
..
HIGH STRENGTH
COST AND THERMAL EXPANSION DISADVANTAGES
DIELECTRIC FILM
DIELECTRIC FILM
PROVIDES HERMETICITY AND DIELECTRIC
BETTER THERMAL EXPANSION MATCH THAN Al
FLAME RESISTANT BACK COVER
USED IN CURRENT PV MODULES
•• COMMERCIAL MATERIAL BEING EVALUATED FOR PV MODULES
•• ·MATERIAL UNDER DEVELOPMENT AND EVALUATION FOR PV
* 0001
In
= 0,025 mm; 0.040 in = 10 mm
75
,170
-
-
-
-
30.00 ·60.00
3.60
1,00 - 2.00
1.00 - 2.00
1.50· 3.00
1.5 ·30
0.7
0.5· 1.0
0.5 . 1.0
0.5 . 1.0
0.075fm 2
0.125
5.50 - 8.50
5.40
10.00
3.0 ·5.0
0.91
1.5
1,00/m 2
1.00
1.00 . 2.00
1.00 - 2.00
0.13
0.05
0.05· 0.13
0.05 - 0.13
-
-
loday's Status
It
It
It
Superstrate encapsulation designs, materials,
and material processing are compatible with
mass production requirements for modules
using crystalline silicon solar cells and should
meet cost allocations.
Superstrate and substrate module design
concepts devised and verified by passing
module qualification tests.
It
Devised photodegradation resistant additives
for pottants and devised and carried out
degradation tests that verified their value.
Encapsulation materials have been life-tested to
20 years.
It
Devised life prediction modeling and tests, and
performed studies of module encapsulation
lifetimes.
It
Encapsulation materials life-tested to 20 years.
It
Materials industry responded very well to JPLdefined material development needs and have
established commercial products.
It
Advanced encapsulation materials are expected
to have a 30-year life expectancy.
It
More testing and field operational results are
needed to determine actual module life
capabilities.
Significant Accomplishments
Encapsulation System R&D Needs
• Assessed module encapsulation requirements
and potential materials for use under terrestrial
operating conditions.
It
Developed a polymeric pottant material (EVA)
and processing techniques suitable for mass
production of PV modules.
It
Developed back-cover concepts and materials.
It
Developed primers and adhesives for glass
superstrate modules using the above materials.
76
It
Thin-film module encapsulation systems.
It
Flexible top cover films.
It
Durable bonding technology for thin-film
modules.
It
Life verification of above and superstrate
module encapsulation systems for crystalline
silicon solar cell modules.
reliability problems a systematic research effort was
undertaken with parallel efforts focused on the most
troublesome failure mechanisms. Initial emphasis,
placed on the development of test methods useful for
quantifying the reliability weaknesses during the
module design phase, led to the continuing evolution
and development of module qualification tests. These
tests and their severity were selected to fail older
module designs with known deficiencies and to pass
modules with good in-the-field performance. As module
designs and fabrication techniques improved, emphasis
was gradually shifted toward achieving a physical
understanding of the longer-term, less-well-understood
failure mechanisms and devising qualification tests and
design solutions for them. Research activities were
initiated studying interconnect fatigue (Figure 69),
sheet glass strength (Figure 70), module soiling (Figure 71), hot-spot heating (Figure 72), breakdown of
electrical insulation materials (Figure 73) and various
others. All of these activities have led to comprehensive amounts of directly usable data and knowledge
as shown in Figure 74. This information has been
quickly transferred to industry and continues to be
incorporated into production module designs and
fabrication techniques.
Reliability Physics
Reliability has always been an inherent part of FSA
activities. In fact, efforts in all parts of the Project have
been directed toward a low-eost, long-life product. As
terrestrial PV technology evolved, the more easy-tosolve failure mechanisms were addressed first; in
parallel, excellent progress was made on long-term
studies of encapsulation materials and designs for longlife modules. As each aspect of module technology
evolved, as an entity, the need to understand all the
degrading mechanisms and their interactions within a
module became apparent. These efforts have been
consolidated and are called Reliability Physics.
The objective is to develop the reliability and
durability knowledge required to produce long-life PV
modules and arrays:
III
Develop an understanding of module long-term
degradation and failure mechanisms and
techniques for reducing degradation.
III
Develop a reliability prediction capability.
III
Develop accelerated test capabilities.
III
Verify material, device, and module reliabilities.
Over the years, the reliability/durability research in
FSA has evolved into a general methodology with six
major elements:
III
Background
Identification of key degradation mechanisms.
• Establishment of mechanism-specific reliability
goals.
In the early 1980s, when the DOE PV Program
emphasis was shifted to electric utility applications,
module and array lifetime goals were increased to
30 years from 20 years. This was considered to be a
practical goal based upon technology advancements,
typical utility planning for 30-year operational times,
and the economic advantages of longer operation of
power systems.
III
Quantification of mechanism parameter
dependencies.
III
Development of degradation prediction methods.
III
Identification of cost-effective solutions.
• Testing and failure analysis of trial solutions.
The longer-life requirement necessitated that a more
basic understanding be determined of all the
phenomena involved in module/array degradation. A
long-range integrated approach of analyses and
experiments covering all important parameters that can
shorten the life and/or reduce the power generating
capabilities of PV arrays was established. Within FSA,
these efforts are now grouped under Reliability Physics
and include the durability of cells, cell metallization, cell
interconnects (within a module), encapsulation
materials and their use, effects on module life of
fabrication processing; and the influence of operational
and environmental conditions.
Although a substantial degree of synergism exists
among these elements, it is useful to address each
separately.
Identification of the failure mechanism is a critical
step. During the early years of the Project, module
qualification testing was effective in detecting module
failure. With the better quality of contemporary
modules, the best evidence of a generic problem is
well documented field failures. This requires careful
monitoring of field applications with statistically
significant numbers of modules, an active problemfailure reporting system, and a comprehensive failure
analysis capability. The voltages and currents in an
operating system also play an important role in hotspot heating failures, shorts to ground and in-circuit
arcs. Failures that show up only after prolonged field
exposure must be identified by a variety of accelerated
tests, unless waiting for years is acceptable.
Project Activities
During the course of the project, a steady stream of
module failure mechanisms (Figure 68) was observed
and identified through module testing, application
experiments, and failure analyses. To resolve the
77
UNiFORMLY LOADED, SIMPLY SUPPORTED,
RECTANGULAR GLASS SHEET
LIFE PREDICTION
OBJECTIVE:
CENTER STRESSES ARE CRITICAL
BELOW BREAK IN CURVES
a:
o
OEVELOP AND VERIFY DEGRADATION RATE
\
~
()
MODELS FOR EACH WEAR-OUT TYPE FAILURE
~
MODULE FAILURE TYPES
1000
>~
u;
WEAR-OUT IAGINGI
- CORROSION
- MATERIAL DEGRADATION
- FATIGUE ICYCLINGI
- WEAR (DIMENSIONALI
- DAMAGE ACCUMULATION
zw
~
~
'"w
'"a:
100
~
'"
(INFANT MORTALITYI
MANUFACTURING
PROCESS QUALITY
FAILURE
RATE
-
10 ~"lL.J...L.L-'-':':'::--!-.l-LL.LllU_.l-L-'-'-LUJ.l_L..L-'-LJ.wJ
10
100
1000
1000
1000
MATERIALS
CONTAMINATION
DIMENSIONS
ASSEMBLY
HANDLING
LOAD INTENSITY FACTOR
I
-
RANDOM
OVERSTRESS
-
RANDOM
FLAWS
TIME IN OPERATION
Figure 68. Module Failure-Rate Classification
Figure 70. Glass Breakage and Sizing Research
TORRANce, CA
PHOTOGRAPHS OF PARTICLES
DEPOSITED ON MODULES
INTERCONNECT
STRAIN-CYCLE FAILURE
TEST DATA
(Scanning Electron Microscope)
~30
z
~
GENEU,UDES'ClNMETHDDOlOC'ES,
TR...lIE-a"o...r .... GUlOu'..E5....NO
g '"
ADVANCED D(s'GNS'Q/I KEY
MOOUlf COMPONENTS lNVOlVING
MOOUlE.o.SSEMDlYlEVElTIIAOE
z
Off$
APPROACH
400 X
... IOENlIfYlMPOIIT.o.NTCOMPONENT
DUIC)N$IIEOUIIII'HI.o.SSEMBlY
lEVElOI'l'IMIZATlO><
.OfVnOPNECE5SAnTlIUI<)N
METHOllOlOGIUANOTfST
nC'''IIOUU
... CONOUCrANAlVSUTOOEfmE
DE$Il)N TnAOE.Qff5. IOfNTIFT
DESl(lNGUIOElINE$AND OEVElOP
AOVA"CEO COMPON[NT DUKlNS
400 X
u
~ 10
Figure 71. Module Soiling Research
Figure 69. Electrical Interconnect Fatigue Research
78
TABLE MY - LOS ANGelES. CA
EXPERIMENTAL APPARATUS
OBSERVED MODULE
RESPONSE vs CELL TEMPERATURE
Figure 72. Hot-Spot Endurance Test Development
Technology
Bond Delamination
Interconnect Fatigue
Galvanic Corrosion
Electrochemical Corrosion
Photothermal Degradation
Structural Failure
Hail Impact
Glass Breakage
Cell Fracture Mechanics
Voltage Breakdown
Hot-spot Heating
Soiling
-MYLAR lEU
A ·04801,1
B
09211111
C
1420111
o
30001,1
-- TEDLAR
E
lOOSG 30 TA, 1 mil
F
100BG30TR,lm,1
G
200 SG 40 TA, 201;1
0.99
HIGH VOLTAGE THIN FILM
INSULATION BREAKDOWN
RESEARCH APPARTUS
THIN FILM
TEST BREAKDOWN DATA
Vear
75
76
77
78
79
80
01
82
C
~
83
84
Cl
Cl
el
e: 1=
1=="
!= 'fOI
=11"
=y
~
e: 13
=1'.
.., = Major Contribution
A
Figure 73. Insulation Breakdown Research
=
Significant Contribution
Figure 74. Reliability and Durability Developments,
1975 to 1984
79
Thirteen principal failure mechanisms of flat
crystalline-silicon modules, grouped into four types of
degradation are shown in Table 22. The key failure
mechanisms identified to date, and target degradation
levels for each that are consistent with a 30-year
lifetime, are shown together with their economic
significances and target allocation levels. The units of
degradation in the third column provide a convenient
means of quantifying failure levels of individual
mechanisms according to their approximate time
dependence. Columns 4 and 5 indicate the level of
degradation for each mechanism that will result in a
10% increase in the cost of delivered energy from a
large PV system. Because the mechanisms will
generally occur concurrently, the total cost impact is
the sum of the 13 cost contributions. Column 6 of the
table carries the analysis a step further and suggests a
strawman allocation of allowable degradation among
the 13 mechanisms to achieve a total reliability
performance consistent with expectations of a 30-year
lifetime. The total effect of the allowable levels is a
20 % increase in the cost of energy over that from a
perfect, failure-free system. The distribution among the
mechanisms reflects the judgment of FSA personnel at
this time. It is important to update these allocations as
field operation experience becomes available in the
next few years.
A specific degradation prediction can be made by
determining the time history of an applied stress of the
appropriate parameter. A variety of these
environmental stress characterizations have been
developed at JPL including hail impact probability,
wind loading pressures, and array voltage and current
durations. Solar radiation surface meteorological
observations (SOLMET) weather data have been used
to model ultraviolet, temperature, and humidity levels
to which modules are exposed. These models have
been combined with other parametric variables in
making module useful-life predictions.
Cost effectiveness of modules and arrays requires
trade-offs of degradation rates, failure rates, and
lifetimes against initial-manufacturing costs, field
maintenance costs, and lost energy revenues. Lifecycle costing analysis integrates these variables and
permits trade-offs to be made. In the identification of
cost-effective solutions, it is important to minimize
sensitivities to processing variations and design and
material uncertainties. These analyses permit rational
judgments and cost-effective levels of reliability to be
selected.
The first and last step in a reliability program is
extensive testing of the complete product. Such testing
cuts across the entire reliability development effort
from beginning to end. It plays a key role in the
identification of failure mechanisms, in the illumination
of parameter dependencies, in the selection of costeffective solutions, and in the verification of final
designs. By the end of the reliability effort, tests with
quantitative correlation to field applications should be
available, based on the parameter dependencies
determined and the prediction algorithms developed. A
key element of complete-product testing is inclusion of
the synergistic reactions between all of the
components, including likely interfacing components on
A specific degradation or failure mechanism is
dependent upon a large number of field-stress
parameters, system operating parameters, componentdesign parameters, and manufacturing-processing
parameters. In order to understand the test data, a
quantitative understanding of the importance of each
principal parameter is required. This is best derived
from understanding the chemical and physical
processes in the degradation, plus a qualitative insight
into the mechanism physics as can best be determined
by specific tests and analyses.
Table 22. Life-Cycle Cost Impacts and Allowable Degradation Levels
Type of
Degradation
Failure Mechanism
Units
of
Degrad.
0.08
0.24
0.05
0.67
0.67
10
0.33
0.33
0.33
0.70
0.022
0.022
0.13
0.40
0.25
1.15
1.15
10
1.18
1.18
1.18
2.40
0.122
0.122
Allocation
for
30·Year·
life
Module
0.005
0.050
0.001
0.20
0.20
3
0.1
0.1
0.1
0.05
0.01
0.01
27
20
35
level for 10%
Energy Cost
Increase*
k
Open-circuit cracked cells
Short-circuit cells
Interconnect open circuits
Cell gradual power loss
Power
Module optical degradation
degradation
Front surface soiling
Module glass breakage
Module open circuits
Module
Module hot-spot failures
failures
Bypass diode failures
Module shorts to ground
Module delamination
life-limiting Encapsulant failure due
wearout
to loss of stabilizers
*k = DIscount rate
%/yr
%/yr
%/yr 2
%/yr
%/yr
%
%/yr
%/yr
%/yr
%/yr
%/yr 2
%/yr 2
Years
of life
Component
failures
80
=0
k
= 10
Economic
Penalty
Energy
Energy
Energy
Energy
Energy
Energy
O&M
O&M
O&M
O&M
O&M
O&M
End of
life
the user's side of the interface. This latter consideration
is one of the reasons field-application data are more
convincing than laboratory test data because they
include interaction with the user.
E!I
E!I
Developed test methods and performance data
on the reliability of solar-eell metallization
systems.
Characterized the electrical breakdown strength
of polymeric dielectric films.
Significant Accomplishments
E!I
E!I
E!I
E!I
E!I
Developed long-term soiling data for various
module surface materials and synthesis of
antisoiling surface treatments.
E!I
Developed analysis tools, test methods, and
design data for the control of solar-eell hot-spot
heating.
E!I
Developed data on probability of impact of
various-sized hailstones, and means of surviving
hail impact.
E!I
Developed analysis tools and design data for
predicting the breaking load of glass sheets.
E!I
E!I
Developed design data and test methods on the
photothermal stability of module encapsulants.
81
Characterized the parameters involved in
electrochemical corrosion of modules.
Developed analysis tools and design data for
predicting interconnect fatigue and designing
long-life, reliable cell interconnections.
Developed design data and qualification testing
technique to ensure the reliability of bypass
diodes.
Developed design data and testing techniques on
the fracture strength of crystalline-silicon wafers
and cells.
Developed a comprehensive set of module
qualification tests, testing facilities, and failure
analysis facilities.
whether module innovations have resulted in
increased efficiency. Therefore, progress would have
been impossible to assess without the development
of precise and accurate methods of performing this
measurement. Because the measurement of modules
under natural sunlight is prohibitively time-consuming,
use is made of solar simulators. Early in the Project,
a standard irradiance (magnitUde and spectral
distribution) was defined and various solar simulators
were put into commercial use by the many
companies and laboratories involved in the program.
Because none of the simulators could duplicate the
standard spectrum, methods were developed for
correcting the simulator measurement to the value of
power output that would have been obtained if the
module had been exposed to a standard irradiance.
odule Performance
and Failure Analysis
In 1975, at the start of the FSA Project, litte or no
quantitative data were available on the electrical
efficiency of the early PV modules or on the exact
cause of the reliability problems common to modules
in use at that time. Building on JPL's history of
making precise measurements and performing failure
analyses as part of its space photovoltaic activities,
the Project initiated the measurements, testing, and
failure analysis of terrestrial modules in 1975. This
was an important part of the Block I module
procurement activities and it has continued
throughout the duration of the Project.
In 1975, electrical performance measurement of
solar cells and arrays had already undergone considerable development in connection with the use of
these devices on spacecraft over the prior 15 years.
Measurement for that application was simpler than
measurement for the terrestrial application because
the solar irradiance spectrum above the Earth's
atmosphere is more stable and more easily measured
and described. It was also feasible to devote
extensive resources to the measurement of a single
spacecraft module (or array) because of its high cost
and need as part of space mission success.
The objectives of the module performance
measurement and failure analysis activity are to
support the module development effort:
.. By providing accurate electrical measurements
of state-of-the-art and advanced development
modules, including developing and providing
international recognized electrical performance
measurements and facilities.
.. By conducting various characterization and
environmental qualification tests including the
development of unique testing methods and
facilities.
In the terrestrial application, the irradiance
spectrum is far more difficult to measure and define
because it is subject to spectral absorption and
scattering in the atmosphere and because it consists
of both a direct and diffuse component. Because the
commercial cost of an individual module must be
low, the resources allotted to its measurement must
also be low.
.. By establishing a problem-failure reporting
procedure, conducting detailed analysis of
module failures, and developing advanced
failure analysis techniques and facilities.
Background
The first method consisted of performing a primary
(natural sunlight) calibration of a solar cell that was a
sample of the cells used in the modules; it could be
presumed to have the same spectral response. Using
this cell as a reference cell for adjusting the
magnitude of irradiance from any simulator, the
resultant measurement was then independent of the
simulator spectrum. Subsequently, a new method
known as mismatch factor correction was developed.
With this method, which required knowledge of the
simulator spectrum and of the spectral response of
both the reference cell and the module, it was
possible to perform a computer correction of the
measured power value even though the reference
cell spectral response did not match that of the
module.
At the start of the JPL Project, it was recognized
that to succeed with the development of advanced
module technologies there must be accurate
measurements data to continually determine the
status of the technology, and to guarantee accurate
communication among the large number of
researchers involved. The need to accurately define
costs and cost standards was the responsibility of
the Project Analysis and Integration portion of the
Project; the need to accurately measure efficiency
and reliability became this activity's responsibility.
This evaluation function served to determine whether
design and fabrication innovations were successful in
their objectives, and to identify deficiencies that
should be the object of design and fabrication
modifications or the subject of further research.
The disadvantage of the above two methods is
that either they assume that the spectral response of
the module is the same as that of the cell, or they
require the difficult or impossible operation of
measuring the spectral response of the module. A
solution to these problems consisted of designing a
simulator/filter combination that produced the
Project Activities
Electrical Performance Measurement
The electrical performance of a new module
design is the necessary criterion for determining
82
standard spectrum. Such a system has now been
implemented for both of the current irradiances
defined in the U.S. standards documents. These are
the airmass 1.5 direct normal irradiance (ASTM E
891-82) and the airmass 1.5 global spectrum
(combining direct and diffuse components, ASTM E
892-82). With this system, the reference cell need
only be of the same generic material as the module's
cells, and very large modules can be measured
because their spectral response need not be known.
180
---- ASTM E891-82 DIRECT IRRAD
- - FILTERED LAPSS DIRECT IRRAD
2
NORMALIZED TO lOOOW/m
160
EI40
N
T
E
~120
g100
,
,I
'E
,,
80
w
u
I
z
~
0
«
The simulator used in this system is known as the
Large Area Pulsed Solar Simulator; it produces the
spectrum shown in Figure 75 when it is not filtered.
The filtered spectra for the direct normal case and
the global case are shown in Figures 76 and 77,
respectively, along with the desired standard spectra.
These spectral matches are close enough in both
cases so that the error due to the mismatch is not
greater than 1%.
'"~
I
60
40
,
I
I
I
I
I
,I
I
20
I
o'
300
400
500
600
700
800
900
1000
WAVELENGTH (nanometers)
1100
1200 1300
Figure 76. Spectrallrradiance (AM 1.5 Direct LAPSS
Versus ASTM AM 1.5 Direct)
180
---- ASTM E892-82 GLOBAL IRRAD
FILTERED LAPSS GLOBAL IRRAD
2
NORMALIZED TO I000W/m
160
J
140
E
~ 120
j
b 100
w
'E
U
z
~
80
t5
z
0
«
'"
~
60
'"'"
---
40
;j
'"
60
/
40
400
500
r---... ........
:
20
300
"
\; I
I
II"
VII
IV
oI
600
700
BOO
900
1000
WA VELENGTH (na nometersl
1100
1200 1300
300
Figure 75. Spectrallrradiance (JPL Unfiltered
LAPSS)
,-
--.I..
\
I
I
20
0
I
o
«
w
>
;::::
80
~
I
400
500
600
700
800
900
1000
WAVELENGTH (nanometersl
1100
1200 1300
Figure 77. Spectrallrradiance (AM 1.5 Global LAPSS
Versus ASTM AM 1.5 Global)
83
Qualification Testing
Figure 78. As the figure also shows (under the
heading "Qualification Test Specifications"), the
specifications evolved during a series of module
development programs over a 10-year period. The
evolution has resulted from the information provided
by the tests themselves, from module field test
experience, and from module research.
In addition to determining whether a new module
design provides a desired increase in efficiency, the
electrical performance measurement of a module
also serves as the principal criterion for determining
the success of a new module in terms of its ability to
withstand the environmental and electrical stresses
expected in the field. An accelerated program of
these stresses has been applied to all new modules
developed under the FSA Project as well as to other
commercial modules available from U.S. and foreign
sources. The test program consists of a set of
formally defined qualification tests whose description
and sequence are illustrated and described in detail in
Qualification tests have been performed on
modules of about 150 different designs. With few
exceptions, the tests have revealed defects
necessitating module design or manufacturing
process changes. This is proof of the critical role of
this type of testing in identifying module
modifications needed, and in identifying required
research in materials, processes, andlor design.
ELECTRICAL PERFORMANCE MEASURED AND PHYSICAL DURABILITY ASSESSED
INITIAL EVALUATION
L...
f'll""
THERMAL CYCLE
TEST
+-
EV
f+.
HUMIOI'iEsrEEZING
~
-
1+
EV
LOADING
EV
v
"SO
elECTRICAL PERfORMANCE
VISUAL AND MeCHANICAL
INSPECTION AND
CHARACTERIZATION
- INSULATION TEST
- NOMINAL OPERATING CELL
TEMPERATURE INOCTI
DETERMINATION
- VOLTAGE AND CURRENT
TEMPERATURE COEFFICIENT
DETERMINATION
~
Ii"'"
... -,. I...
1-pIi'
M'CHAN'CAC
-
_
a...._ _--I
HAil
I~
...
a...._'_M._m_...I
FINAL
EVALUATION
ELeCTRICAL
PERFORMANCE
VISUAllNSPECTtON
INSULATION TEST
EVALUATION OF
CHANGES
.....
CYCLIC PRESSURE LOADING TEST APPARATUS
ElECTnlCAl PERFORMANCE
HAIL TEST APPARATUS
VISUAL INSPECTION
TYPICAL THERMAL CYCLING TEST
I------ MAXIMUM
CYCLE TIME -
HUMIDITY-FREEZING CYCLE
..I
:,
ICONTINUED FOR
l/lO"CIh,
MAXIMUM
5(I~1~~
.
:'\
, ,
j
'-
__
...J
,/
I
TIME(hr)
QUALIFICATION TEST SPECIFICATIONS
<O"<;YO'!IO'C,CYCUSAS,"O'CATlO
'0
~CVCUSAY""~H,n-c~o..,.c Oil
\DCYClESMlIS.. m'.,oIO'CTO_lIS'C
MICHA,,'CAl
lOAO'''GCYClE
~~~~~i~!>O 'O-'f,2l CYCLES
AS
,~~:~:,'::~~~,~ lA8 ~ur ..o "7
~t'.:'G~~""Ell"'TEII2<""'m."'OF
2O ........ 0'A... ETEnHAIL·21m1<
:IOOll
LARGE-AREA
SOLAR SIMULATOR
TESTING
!>O_AMAX CURIIE"TATVOLTAGE
W ........".... '.." .. 'MUM~ESISTA.. C[TO
G~OU.. Ofonu~OSEOCO
.. OIlCmnS
i~'!~Z~\',~::'~~'t.:;,::~~:,
• liES _ ~ESJOE ..T!AL
,,- '''1E~MEOIATE lOAO
Figure 78. Module Qualification
84
Qualification tests are performed on a small
sample of modules of any given design. The proof of
satisfactory performance, including reliability,
requires field testing the modules in the various
environments in which PV power stations would be
installed. Such a program was instituted early during
the Project history using the sites listed in Figure 79.
The performance observed in the tests at these sites,
and at other array installations that were part of the
DOE program, resulted in important modifications to
the earlier qualification test specifications. These
provided for better accelerated testing to reveal the
potentiality of failures because of such phenomena
as hail impact, thermal stress, and cell back-biasing.
The early history of failure statistics is shown in
Figure 80. Because of reductions in support of field
testing, statistics are not available for the Block IV or
Block V modules.
perform an in-depth failure analysis to find the exact
cause of the problem. A Problem Failure Reporting
system was established by the FSA Project in 1975
to provide formal reporting of all failures, regardless
of the site of occurrence. The reports and the failed
modules were delivered to failure analysis personnel
who then applied a variety of sophisticated techniques (some derived from the NASA space exploration program and some developed explicitly for the
terrestrial PV program) to isolate the specific cause of
the failure (Figure 81). A highly detailed report was
prepared describing each such analysis and
presenting the results, with recommendations for
correcting the module deficiencies. These
Performance and Failure Analysis Reports were
supplied to the manufacturer of the module and to
JPL personnel responsible for module development
research. During the various module development
programs, such analyses were instrumental in
correcting design and processing problems to the
extent that new modules incorporating the
recommended changes were then successful in
passing the qualification tests.
Failure Analysis
When problems occur during qualification tests, or
field tests, or at array installations, it is necessary to
MODULE FIELDTESTING-16 SITES
FIELD TEST SITES
EXTREME WEATHER
FORT GREELY. ALASKA (ARCTIC)
FORT CLAYTON. CANAL ZONE (TROPIC)
MARINE
KEY WEST. FLORIDA
SAN NICOLAS ISLAND. CALIFORNIA
POINT VICENTE. CALIFORNIA
HIGH DESERT
ALBUQUERQUE. NEW MEXICO
DUGWAY. UTAH
GOLDSTONE. CALIFORNIA
PASADENA. CA IJPLI
PRIME SITE FOR FIELD
TESTING
MOUNTAIN
MINES PEAK. COLORADO
TABLE MOUNTAIN. CALIFORNIA
URBAN COASTAL
NEW ORLEANS. LOUISIANA
NEW LONDON. CONNECTICUT
MIDWEST
CRANE. INDIANA
UPPER GREAT LAKES
HOUGHTON. MICHIGAN
NORTHWEST
SEATILE. WASHINGTON
URBAN SOUTHWEST
PASADENA. CALIFORNIA
HOUGHTON. MI
TYPICAL REMOTE SITE
Figure 79. Module Field Testing (16 Sites)
85
NEEDS FOR LONG-LIFE, DURABLE MODULES/ARRAYS
• MODULE AND ARRAY FIELD TESTING
• REQUIRED BECAUSE MODULE QUALIFICATION TESTING HAS NOT BEEN ENTIRELY
SUCCESSFUL IN PREDICTING AND/OR DUPLICATING MODULE FIELD FAILURES
• FIELD TESTING IN REAL-USE POWER
SYSTEM CONFIGURATIONS
• DETECTION OF DEFICIENCIES/FAILURES
WHICH OFTEN HAVE BEEN A SURPRISE
• KNOWLEDGE OF MODULE WEAR-OUT MECHANISMS WHICH, OTHER THAN CELL
INTERCONNECT FAILURES, HAVE NOT BEEN DETECTED AS YET; POTENTIAL
LONG- TERM OPERATION DEFICIENCIES/FAILURES ARE EVIDENT (CRACKED CELLS
AND COVER GLASS)
• PERFORMANCE VERIFICATION, UNDER
OPERATIONAL CONDITIONS, OF NEW
ADVANCED MODULES, WITH INNOVATIVE
DESIGN FEATURES
• REDUCTION OF SINGLE MODULE ENDURANCE FIELD TESTING IN FAVOR OF
MODULE TESTING IN ARRAYS
• KNOWLEDGE OF MODULE/ARRAY
PERFORMANCE AND SAFETY PROBLEMS
WITH THE REMAINDER OF THE
POWER SYSTEM AND WITH
UTILITY INTERACTIONS
FAILURE" RATES FOR MODULES AT FSA FIELD TEST SITES
21
0
261 MODULES DEPLOYED
48 FAILURES
UJ
..J
<
u..
In
UJ
..J
15
0
0
12
:::>
:E
• CENTRAL STATION: ESTABLISHMENT
OF CREDIBILITY OF MODULE/ARRAYS
AS POWER GENERATORS FOR
CENTRAL STATION USE; i.e.,
COST EFFECTIVE WITH
LONG LIFE OPERATION
18
BLOCK I
u..
0
UJ
C)
<
IZ
9
326 MODULES DEPLOYED
12 FAILURES
6
UJ
(.)
a::
3
UJ
a..
1
24
63 MODULES DEPLOYED
o FAILURES
30
36
42
48
·FAILURE - OUTPOWER DECREASE
GREATER THAN 25%
TIME IN FIELD, MONTHS
Figure 80. Needs for Long-Life, Durable Modules/Arrays
MODULE PROBLEM/FAILURE ANALYSIS
LASER SCAN OF MODULE
OUTPUT SHOWING CRACKED CELL
X-RAY OF OVERHEATED
CELLS SHOWING MELTED SOLDER
HIGH VOLTAGE
BREAKDOWN
OF INSULATOR
PHOTOMICROGRAPH OF HAIL DAMAGE
CELL INTERCONNECT FAILURE
• PROBLEMS/FAILURES AT TEST/APPLICATION
SITES REPORTED
• JPL AND MANUFACTURER EVALUATE
P/F AND DETERMINE CORRECTIVE ACTIONS
• MANUFACTURER CHANGES MODULE
DESIGN OR WORKMANSHIP AS NEEDED
Figure 81. Module Problem/Failure Analysis
86
• Developed module inspection techniques and
guidelines.
Significant Accomplishments
• Developed a unique Large Area Pulsed Solar
Simulator with closely matched airmass 1.5
direct and airmass 1.5 global spectra.
• Environmentally tested more than 150 different
module designs to date over the 10-year period.
• Processed 1200 reports of failures. involving
460 major failure analyses.
• Calibrated numerous reference cells that have
served as the primary industry standards over
the past 10 years.
• Developed unique environmental testing
facilities including hail guns, mechanical cyclicloading apparatus, hot-spot test equipment. and
nominal operating cell temperature (NOCT) test
racks and instrumentation.
• Developed a simple. accurate method of
secondary calibration of reference cells; have
calibrated cells for many U.S. and foreign PV
manufacturers.
• Established outdoor field test-rack facilities
throughout the United States in various
geographical climates.
•
87
Developed a unique laser scanner instrument
for the failure analysis of large-area moduLes.
Appendix A: Glossary
A
Ampere(s); Angstrom(s)
A&E
Architects and engineering
AM
Air mass (e.g., AM1
AR
Antireflective
a-Si
Amorphous silicon
ASME
American Society of Mechanical Engineers
ASTM
American Society for Testing and Materials
BSF
Back-surface field
BSR
Back-surface reflection
C
Celsius (temperature scale)
CAST
Capillary Action Shaping Technique
COSMIC
Computer Software Management Information Center
CVD
Chemical vapor deposition
CVT
Chemical vapor transport
Cz
Czochralski (classical silicon crystal growth method)
DCS
Dichlorosilane
DOE
U.S. Department of Energy
EFG
Edge-defined film-fed growth (silicon-ribbon growth method)
EMA
Ethylene methyl acrylate
EPSDU
Experimental process system development unit
ERDA
Energy Research and Development Administration
EVA
Ethylene vinyl acetate
FAST
Fixed Abrasive Slicing Technique
FBR
Fluidized-bed reactor
FSA
Flat-Plate Solar Array Project
FZ
Float-zone (silicon crystal growth method)
h
Heat transfer coefficient; hour(s)
HEM
Heat-exchange method (silicon-crystal ingot-growth method)
Isc
Short-circuit current
I-V
Current-voltage
ID
Inside diameter
= unit air mass)
88
IPEG
Improved Price Estimation Guidelines
IR
Infrared
J sc
Short-circuit current density
JPL
Jet Propulsion Laboratory
Kg
Kilogram
LAPSS
Large-area pulsed solar simulator
LASS
Low-angle silicon sheet growth method
MEPSDU
Module experimental process system development unit
mg-Si
Metallurgical-grade silicon
MINP
Metal insulator, n-p
MOD
Metallo-Organic deposition
MSEC
Mobil Solar Energy Corp.
MT
Metric ton(s)
NASA
National Aeronautics and Space Administration
NEC
National Electrical Code
NOCT
Nominal operating cell temperature
O&M
Operation and maintenance
Pmax
Maximum power
PA&I
Project Analysis and Integration Area (of FSA)
PDU
Process development unit
PIM
Project Integration Meeting
ppba
Parts per billion, atomic
ppm
Parts per million
PU
Polyurethane
PV
Photovoltaic(s)
PVB
Polyvinyl butyral
R&D
Research and development
RH
Relative humidity
s
Second(s)
SAMICS
Solar Array Manufacturing Industry Costing Standards
SAMIS
Standard Assembly-Line Manufacturing Industry Simulation
SCAP1D
Solar-Cell Analysis Program in One Dimension
89
SCCD
Short-circuit current-decay
SEEMA
Solar-Cell Efficiency Estimation Methodology and Analysis
SERI
Solar Energy Research Institute
Si
Silicon
SIMRAND
SIMulation of Research And Development Projects
SMUD
Sacramento (California) Municipal Utility District
SOA
State of the art
SOC
Silicon-on-ceramic
SOLMET
Solar radiation surface meteorological observations
STC
Silicon tetrachloride
T
Temperature
TCM
Transparent conducting material
TCS
Trichlorosilane
UL
Underwriters Laboratories, Inc.
V oc
Open-circuit voltage
Wp
Peak watt(s)
90
Appendix B: FSA Project Contractors
COMPANY
co
AeroChel'l Research Laboratories, Inc.
AeroChel'l Research Laboratories, Inc.
AeroChePl Research Laboratories, Inc.
Aerochel'l Research Laboratories, Inc.
AlA Research Corp.
Applied Solar Energy Corp.
Applied Solar Energy COrp.
Applied Solar Energy Corp.
Applied Solar Energy Corp.
Applied Solar Energy Corp.
Applied Solar Energy COrp.
Applied Solar Energy COrp.
Applied Solar Energy COrp.
Applied Solar Energy Corp.
Applied Solar Energy COrp.
Applied Solar Energy Corp.
ARGO Solar, Inc.
ARGO Solar, Inc.
ARGO Solar, Inc.
ARGO Solar. Inc.
ARGO Solar, Inc.
ARGO Solar, Inc.
ARGO Solar. Inc.
AriZONl State University
Astrosystel'lS, Inc.
Battelle I1el'lOrial Institute
Battelle I1el'lOrial Institute
Battelle I1el'lOrial Institute
Battelle I1el'lOrial Institute
Bechtel National, Inc.
Bechtel NationaL Inc.
Bernd Ross Associates
Bernd Ross Associates
Burt Hill Kosar RittelAann Associates
Burt Hill Kosar RittelAann Associates
Burt Hill Kosar RittelAann Associates
C. T. Sah Associates
C. T. sah Associates
California Institute of Tecmology
california Institute of Tecmology
Carnegie Mellon UrQllersity
Case IJestern Reserve i.X1iversi ty
Cttronar Corp.
Clel'lson University
Coors Porcelain Co.
Cornell University
Cornell University
Crystal SystePlS. Inc.
COMTRACT TITLE
Si Halide-Al"'.ali. I1etal FlaPles as a Source of Solar Grade Si
Developl'lent Processes for the Production of Solar Grade Si frm Si Halides [; Alkali Metals
Developl'lent of tlodel and COl'lPuter Code for Si Reactions
Synttlesis of Silane and Si in a Non-E~ilibrim Plasna .Jet
Integrated Residential PV Array Developl'l€nt
Developl'lent of Hi!1J-Efficiency Solar cells
Evaluation of Io~n-I'~lanted Solar Cells
Developl'lent of Hi!1J-Efficiency (14%) Solar cell Array Modules
Assessl'lent of Present State-of-the-Art SaloJing Tech. of Large Dial'leter Ingots for Solar Sheet l1aterial
High-Efficiency, Lor~-Life Terrestrial Solar ParEl
Design, fi.tlrication, Test and Price ANllysis of Third Generation Desi!)'l Solar Cell l'tldUle
Developl'lent of Low-Cost CCfltacts to 5i Solar cells
Laboratory services
Microcrystalline Si Growth for Heterojunction for Solar Cells
Interl'ledlate Load Nodules for Test and Evaluation
Solar Cell Process DeveloJ)'lent
Solar Cell Panel Develo~ent Effort
VacUUPl Die cast of Si Sheet for PV Applications
Design, fi.tlricatiCfl, Test. lNalification and Price Pnalysis of Third Generation Design Solar cell Modules
Autmated Solar Panel Assel'1bly Line
Design of Block V Solar Cell Modules - 19B1
Block V Docmentation and Solar cell Modules
Adapt Pulsed Excil'ler Laser Processing Tecmiques to fabricate Cost-Effective Solar cells
Conduct an Evaluation and Calibration of a Czochralski (Cz) Crystal Growth Systel'l
Si filA Solar Cell Process
Part 1. - Evaluation of Si Production SUiti.tlle for Solar Cells
Nodule Encapsulation Task of Low-Cost Si Solar Array Project
Study ProgIaPl to Develop 1; Evaluate Die am Container Materials
Study ProgIaPl to Develop 1; Evaluate Die am Container Materials
Terrestrial Central Power Utility Array Life-Cycle Pnalysis
Developl'lent of ReQirel'lents for Terrestrial Solar Arrays
Developl'l€nt of an All Metal Thick fill'l Cost Effectille Metallization S,lstel'l for Solar Cells
Econol'lical Il'Iproved Thick fill'l Solar Cell Contact
Operation and Maintenance Cost Data for Residential Photovoltaic Nodules/Panels
COAl'lerciaVIndustrial PV l'tldules Requirerlents Study
Residential PV Modules ReQjireAents Study
A St.udy of Effect of Il'lpurities in Si Material
A Study of Relationships of Material Properties and Hi-Eff Solar cell Perfomance on Material COl'IPOsition
InvestioatiCfl of free Space Reactor Si Production
Diffusion Barrier Studies
Exploratory Study of Product Safety and Product Liability Considerations for Photovoltaic Module/Array Devices
Encapsulation Syst€l'l Studies for Low Cost Si Solar Array
Al'lorphous Si Module Research
Investigation of Reliability Attributes am Accelerated Stress factors on Terrestrial Solaz Cells
Study PrograPl to Develop and Evaluate Substrate and CCfltaiJler Materials
Investigation of Physical Structure and the Chenical Nature of Defects in Si $t,eet Material
Cheracterization of StructuraL Electrical E; Chel'lical Properties of 5i Sheet t1aterial
Nul ti-Wire Wafering Technology DeveloPRent by a flx€':l Abrasive Slicing Tedmique (f~lST)
OP~
.
.
..
AREA
SIMAT
SINAT
SINAT
SIMAT
RES
DR
PO
PO
PO
PO
MPfA
PO
PO
OR
.
.
.
.
.
..
..
..
..
..
RES
PO
PO
SINAT
MPFA
PO
MPFA
RES
PO
SINAT
SHEET
SINAT
SINAT
SHEET
SHEET
PA6I
RES
PO
PO
RES
RES
RES
SINAT
OR
SINAT
PO
RES
DR
RES
RES
SIMAT
PO
SINAT
SINAT
CDI1PANY
en
I\)
Crystal Systens, Inc.
000 Corning Corp.
000 Corning Corp.
Eagle-ficher Ind.. Inc.
ElectriN<., Inc.
Endurex Corp.
Energy I1aterials Corp.
Energy Materials Corp.
Energy Materials Corp.
Energy Materials Corp.
General Electric Co.
General Electric Co.
Gffleral Electric Co.
General Electric Co.
General Electric Co.
General Electric Co.
Gnostic Concepts. Inc.
Henlock Seniconductor Corp.
Honeywell. Inc.
IBN Corp.
lIT Research Institute
Illinois Tool loklrl<s. Inc.
J. C. SCtIJl'Iacher CO.
Ka'jex Corp.
Ka'jex Corp.
Ka'jex Corp.
Kinetic Coatings. Inc.
Kulicl<e & Soffa Industries
Lanar Uni versi t.y
Locl<heed t1issiles & Space Co.
Locl<heed t1issiles & Space Co.
Locl<heed tti.ssiles & Space Co.
Locl<heed t1issiles & Space Co.
117 International. Inc.
N7 International. Inc.
Nas~chus~tsI~tituteofkmroM~
Nas~chus~ts Institute of Temrolo~
Naterials Research. Inc.
Naterials Research. Inc.
Naterials Research. Inc.
tti.tre Corp.
MOOil Solar Energy Corp.
MOOU Solar Energy Corp.
MOOil Solar Energy Corp.
MOOU Solar Energy Corp.
MOOil Solar Energy Corp.
NCil~nto Research Corp.
Notorola. Inc.
CONTRACT TITLE
Heat-Excha1ger- Ingot casting/Slicing Process, Si Sheet Growth
Task I - Si Naterial Part 3
Develop Si Encapsulation $yS'tens for Terrestrial Silicon Solar Arrays
Study Progrart to Develop and Evaluate Die aa1d Container Materials
Laboratory services for the Devel. & Charac. of Screenable Mats. Utilized in the Fabrication of Solar Cells
Solar Cell Nodule lCil-flating Process and Testing
Electrochenical Production of Si
Low-Angle Si Sheet Growth
Gaseous Nel t Replenishlfflt Systen
Technology Develo~t for LOIol-Angle Si Sheet (LASS) Growth Technique for PI) Applications
Developnent & Testing of Shingle-Type Solar Cell Modules
Integrated Residential Phatovoltaic Array Developnent
Design, fabricatiCil, Test, Qualification and Price Analysis of Third Generation Design Solar ceu Nodules
Design of Block V Solar Cell Nodule - 1981
Floating SUbstrate Sheet Growth Process LASS Task for LSSA Project
Bypass Diode Encapsulation Study
Industrialization Stuctj
Developnent of a Polysilicon Process Based on Chem.cal lJa~r Depositicn
Dip coating Process. Si Sheet GrOlYth DeveloPf'lent for the Large Area SUiCCfl Sheet Task
Si Rit:tJCil Grolorth Using a capillary Shaping Technique
TechniCal St.pport in Reliability Engineering of Photovoltaic Modules
DeRonstration of Gapabili t)I t.o Metallize Solar cells by ICfl-flating
High velocity ContiNJoos-Flow Reactor for Production of Solar Grade Si
Low-Cost CZochralski Crystal Growing Techmlogy
Developnent of Advanced Nethods for CCiltiNJOOS C-zochIalski Grwth
AdIIa1ced CZ Grolorth Process to Produce LC4I-Cost 150 Kg Si Ingots frOl'l a Single Crucible for Techmlogy Readiness
Phase 2 of the Array AutoPlated Asseroly Task
Autonated Solar NOOule AssePlbly Line
Process feasibility Stuliy in SUpport of Si Material
Spraylon Fluorocarbon Encapsulation for Si Solar cell Arrays
Transparent Solar Cell Mowle Developnent Effort
Phase 2 of the Array Autol'lated Asseroly Task
Evaluation of Laser Amealing of Solar cells
Five (5) Kilowatt Solar Cell Nodules
Three (3) KilC41atts of Solar cell Nodules
Hi Speed Growth of Si Sheets in Inclined Meniscus Configurations
Investigation of the I+jdrcgenatiCil of SiCl4
Analysis of Defect Structure in Si
Quantitative Analysis of Defects in Si
Analysis of Defect Structure in Si
tti.tre Solar Energy Test Systel'l Evaluation
Solar Modules
Design of Block V Solar Cell Modules - 19B1
Edge-Defined filn fed Growth for Si Growth Devel~fflt
Grys:ta1 Growth by Edg~Defined filn-fed Growth TechniQJe (EfG)
Block V Doctnentation and Solar cell Hodules
Si Material Task - Part 2
PV Module Electric.al Tel1'lination 0esif11 Requirenents Study
0lPEJI
AREA
SHEET
SINAT
..
..
.
PO
SHEET
PO
PO
SINAT
SHEET
SINAT
SHEET
PO
RES
NPfA
NPFA
SHEET
..
RES
PA&I
SINAT
SHEET
SHEET
RES
PO
SINAT
SINAT
SINAT
SINAl
PO
PO
SINAl
RES
PO
PO
PO
..
..
NPfA
NPfS
SHEET
SINAT
SINAT
SINAT
OR
PO
PO
NPfA
SHEET
SHEET
RES
SINAT
RES
COMPANY
<0
0J
Motorola, loc.
Motomla, loc.
Motorola, loc.
Motorola, loc.
Motorola, loc.
Motorola. loc.
Motorola, loc.
Motorola, loc.
Motorola, loc.
Motorola, Inc.
Motorola, loc.
Motorola, loc.
Motorola, loc.
Nat'l Research for Geosciences Labs, Inc.
North Carolina AT&T Fourdation, loc.
North Carolina State Lhiversity
Northrop Corp.
Oregon State Lhiversi ty
P. R. I-«lffl'lan (tt>rlin Inl1Js.)
P. R. I-«lffPlan (tt>rlin Inl1Js.)
Pennsylvania State Lhiversi ty
Photowatt Intl,Inc.
Photowatt Intl,Inc.
Photowatt Intl.loc.
Photowatt Intl.Ioc.
Photowatt Intl. Inc.
Photowatt Intl.lnc.
Photowatt Intl.lnc.
Photowatt.Intl.Inc.
Photowatt. Intl. Inc.
PolytectYlic Institute of New Yorl< - Albany
Purl1Je Research Fourdation
Purl1Je Research Foundation
RCA Corp.
RCA Corp.
RCA Corp.
RCA Corp.
RCA Corp.
RCA Corp.
RCA Corp.
RCA Corp.
Research Triangle lnsti 11Jte
Rocl<\4ell International Corp.
Rocl<\4ell International Corp.
Rocl<\4ell International Corp.
~mingOectronAna~~s~oouw~
SCience Applications. Ioc.
Silicon Tectnology Corp.
CONTRACT TITLE
Design. Fcbricate. Test, QJalification Eo Price Analysis of 3rd Generation Design Solar cell Nodules
Si Material Task. Part 1
Processing ExperiPlents on the non-Cz Si Sheet
EstalllishPlent of a Production-Re.aay Nfg. Process Utilizing Thin-Si Substrates for Solar Cells
Phase 2 of the Array Autonated Assel'lbly Task
Metallization of Large Si Yafers
Solar Cell J1Ol1Jles with Parallel-oriented lntercomections
Studies ard Testing of Anti-Reflective Coatings for Soda-LiPle Glass
Devel~Plent of a I1ethod of Producing Etch Aesistant Yax Patterns on Solar Cells
Laser-Zone Growth in a Ritbon-to-fUbbon Process (ATA), Si Sheet Growth De'lelopAent
Anti-Reflective Coatings on Large Area lAlass Sheets
Anti-Reflective Coatings ~plied by Acid-Leaching Process
Phase I of the Autonated Array Assel'lbly Task
Evaluation of LSA Test and Perfornaoce I1ethods
Study of Effects of Row-to-Roll Shading on PV Arrays
Rapid Themal Processing of Ion IFllla-t:ed Si
Lifetifle l1easurel'lent Analysis
Devel~Plent of Fluidized-Bed 5i Tec~JIlology
Slicing of Single Crystal {'; Polycrystalline 5i Ingots usifl;l Nulti-fllade Saws
Nul tiple-Blade sawing <t1lS) of Si Ingots
Use of Low Energy Hydrogen Ion lAPlants in Hi-Efficieocy CI'jStalline Si Solar Cells
Phase 2 of the Array Autonated i\ssel'lbly Task
Devel~Plent of Technique for All Coatings Eo Nickel ('; C~per Metallization of Solar Cells
InterPlediate Load Nodule for Test & Evaluation
In-Depth Study of Si LJafer SUrface Te;<turing
Devel~Plent of Low-Cost Pol'jSilicon Solar Cells
4) Kilowatts Solar cell Nodules
Eight Kilowatts of Solar cell Nodules
Devel~Plent of low-Cost. Hi -Energy/lilit-Area Solar cell Modules
Design. Fabricate. Test. QJalification Eo Price Analysis of 3rd Generatioo Design Solar cell Nodules
Devel~Plent & DeI'lonstratioo of S)lnthetic Procedures for Polyl'leric UV Stabilizers & Absorbers
IrK Jet Printing of Silver Metallization for PV
MOO Silver I1etallization for PV
Epitaxial Si Growtll for Solar Cells
Ptlase I of the Automted Array AsseI'lbly Task
Phase 2 of the Array Autonated Assel'lbly Task
Inverted Stepanov Tech (1ST) for Si Sheet Growth DevelopAent
Devel~Plent of I1egasonic Cleaning for Si IoJafers
Evaluation & Verification of Epitaxial Process sequence for Si SOlar cell PromJction
Design of Bloc!< II Solar Cell l1Ol1J1es - '9B1
Develop. & Evaluation Die Nats. for Prol1Jcing Si Ribbons by the Inverted Ribbon Growth Process
CMprehenisve Si Solar cell COA~er tlodeling
Devel~ & DeI'Ionstrate l'1etrodS for Accelerated & Abbreviated Testing of Encapsulated Materials
ChePlical Vapor Deposition Growth Si Sheet Growth Develo~ent
Study Progran for Encapsulation Materials Interface for Low-Cost Si Solar Array
Material Itlalysis of P\I Cells and Nodule Test SaAj:lles
Analysis of Cost-Effective PV Panel Design Coocepts Usifl;l Light Trapping
Advanced IO Yafering Techrology
OPEII
AREA
NPFA
SiNAT
PO
PO
PO
PO
PO
PO
PO
SHEET
SHEET
SHEET
m
..
..
l'PFA
RES
OR
SINAT
SINAT
.
..
..
..
.
SHEET
SHEET
m
PO
PO
RES
PO
PO
RES
RES
PO
l'PFA
RES
PO
PO
SHEET
PO
PO
SHEET
..
PO
PO
l'PFA
SINAT
m
PO
SHEET
..
PO
RES
PO
SHEET
COMPANY
<0
-l'o
Silicoo Tecmology Corp.
Siltee Corp.
Siltee Corp.
Solar Energy Research Institute
Solar Energy Research Institute
Solar Energy Research Institute
Solar Pouer Corp.
Solar TecmolO'y International
Solarelectronics. Inc.
Solarex Corp.
Solarex Corp.
Solarex Corp.
Solarex Corp.
Solarex Corp.
Solare>: Corp.
Solarex Corp.
Solarex Corp.
Solarex Corp.
Solare>: Corp.
Solarex Corp.
Solarex Corp.
Solavolt International
Solec International. Inc.
Solenergy Corp.
Sollos. Inc.
Sollos. Inc.
Speetrolab. Inc.
Spectrolab. Inc.
Spectrolab, Inc.
Spectrolab, Inc.
Speetrolab, Inc.
Spectrolab. Inc.
$peetrolab. Inc.
Spectrolab. Inc.
Spire Corp.
Spire Corp.
Spire Corp.
Spire Corp.
Spire Corp.
Spire Corp.
Spire Corp.
Spire Corp.
Spire Corp.
Spire Corp.
Spire Corp.
Spire Corp.
Spire Corp.
Spire Corp. (SlAulated Pl)lsics. Inc. )
~TRACT
TITLE
of Methods of Prowcing LASS tJy the Slicing of Silicon Ingots Using 10 saws
Develqll'lent of Advanced Methods for CootirlJous Cz Growth
Enhanced ID Slicing Technology for Silicon Ingots
Purification of Metallurgical Grade Si tJy CheFlical Vapor Transport
Solidll1elt Interface Studies of High-Speed S:i Sheet Growth
Optinization of Si Crystals for Hi-Efficiency Si Solar cells
Design. FcDricate. Test. Q.lalification I; Price Analysis of 3rd Generatioo Design Solar cell Nodules
Solar Cell Panel Developllent Effort
Hydrochlorination of SiCl4 I; Metallurgical Grade Si
Process Research of Polycrystalline Si Haterial
Block V DocU'lentatioo and Solar cell !'Mules
Design of Block V Solar Cell I10Wles - 19B1
Process Research of Polycrystalline Si Haterial
Phase 2 of the Array Autol'lated Assevlbly Task
Analysis of the Effects of IApurities on Si
Si Solar cell Noduless with a Total PQI,Ier Capability of 30 Kilowatts
Tecmical feasibility & Effective Cost of Various Wafer Thicknesses for Manufacturing
Higr,..Efficiency, High-Density Terrestrial Solar Panels
Solar Breeder feasibility Investigatioo
Design. fabricate. Test. Q.lalification I; Price Analysis of 3rd Generatioo Design Solar cell Nodules
Nitre Solar Energy Test SystePl E'Yaluatioo
InterPlediate Load Nodules for Test and Evaluation
Aotireflecti'Ye Coatings to Single Crystal or Polycrystalline Solar cells ~ Use of Roller Coating Nethod
InterPlediate Load Nodules for Test and E'Yaluatioo
A Nell Nethod of Netallizaticn for Si Solar cells
Investigation of Nickel Silicon Hetallization Process
High Resolution, Low Cost Solar cell COntact DeveloDAent
Phase 2 of the Array Autol'lated Assef'lbly Task
Design, Oe....elopPlent & Prototype Production of Si Solar cell Nodules
High Resolution. Low-Cost Solar cell Contact DeveloDAent
Analysis of Effects of IJIlpurities Intent-ionally Incorporated into 5i
Si Solar cell Process OevelopPlen'l Fabrication and Analysis
llevelopPlent of Metallization Process
Design, Analysis, Ii Test verification of Advanced EncapsulatiCil System
DevelopPlent & fabrication of a SOlar cell JlTIction Process SysteA
HerPletic Edge Sealing of PIJ Nodules
Ion I~lantation of Non-Cz Si
verification & IApleAentation of ElectrCaStatic Bonding Technologies
Design of Bloev. V SOlar Cell I10Wles - 19B1
Block V OocUAentation and Solar Gell ItldUles
Adapt Pulse E:<ciner Laser Processing Technology to Fabricate Cost-Effective Solar Cells
High-Efficiency Solar Cell I10dules
DevelopPlent of Pulsed Processes for the Manufacture of Solar cells
Design. fabricate, Test. Q.lalification & Price Analysis of 3rd Generation Design Solar cell Nodules
Develc~:lI'Ient
DevelopandTestmca~wation~terials
Solar Cell Panel Developllent
Assess Governflent ().ned JPL Advanced CZO crystal GrOUtfl SystePl l10del
Hi-Rate Lou-Energy Expenditure fabrication of Solar Cells
OPEM
.
..
.
..
.
..
.
..
.
..
..
.."
..
AREA
SHEET
SHEET
SHEET
SINAT
SHEET
SHEET
NPFA
PO
SIMAT
PO
RES
NPFA
PO
PO
SIMAT
RES
PO
PO
PO
NPfA
NPfA
RES
PO
RES
PO
PO
PO
PO
NPFA
PO
SIMAT
SINAT
PO
RES
PO
PO
PO
RES
NPfA
RES
PO
RES
PO
NPfA
PO
PO
SHEET
PO
COtITRACT TI TLE
COI1PI\H't'
<0
(Jl
Springllorn Labs, Inc.
Springllorn Labs, Inc.
SRI International
Stanford University
Stanford University
State Universi ty of New Yorl<-ll1.barYi
$l.Perwave TeclYlOlogy, Ire.
$l.Perwave TeclYlOlogy, Ire.
Texas Instr~ents, Inc.
Texas Instnl'lents, Inc.
Texas InstrU1ents, Inc.
Texas InstrU1ents. Inc.
Texas Instr~ents, Inc.
Texas InstrU1ents. Inc.
Texas Research & EnlJineering Institute
The Aerospace Corp.
The Boeing Co.
The Nitre Corp.
Theodore Barry & Associates
Theodore Barry & Associates
Theodore Barry & Associates
Tracor 113 Associates
Tracor 113 Associates
Tracor 113 Associates
Tracor 113 Associates
Tyl~ Corp.
lk'iderwriters Labs. Inc.
lk'iderwriters Labs. Inc.
Lhioo Carbide Corp.
Lhioo Carbide Corp.
Lhiversity of california at Los Angeles
Lhiversity of california at Los Angeles
Lhiversity of california at Los Angeles
Lhiversity of california at Los Angeles
Lhiversity of california at Los Angeles
Lhiversity of california at SlJ'lta Barbara
Lhiversity of california at SlJ'lta Cruz
Lhiversity of Florida
university of Illinois
Lhiversity of Kentucl<.y
Lhiversity of Massachusetts
Lhiversity of I'lissouri
Lhiversity of Pennsylvania
Lhiversity of Pennsylvania
Lhiversity of Pennsylvania
Lhiversity of Pennsylvania
Lhiversity of Pennsylvania
Lhiversity of SOuth Carolina
Nodule Encapsulation Task
Fabricaticn of Solar cell Nodules
Novel O[.plex ValXlr-ElectrochePlical Method for Si Solar cells
Study of ParaYleters of Heavily Doped Si
Interfacial Barriers in Hi-Efficiency Crystalline Si Solar cells
OXygen & carbcn-Related Defects on Hi-Efficierq Si Solar Cells
Nicrowa'/e Heating in a Non-Reactive FBR
Depositing 5emconductor Layers Using Microwa'ie Tec~f1iques
Si Material Task 1. Part 3
Closed Cycle Process for Low-Cost Solar Si Using a P.otary Chaffier Reactor
Si Crystal Gr0\4th & Waferino Process Il'Iprovef4ents
Phase 2 of the Array AYtol'lated Asseroly Task
Adv~ced Methods for Continuous Cz GrOYth, Si Sheet Growth Developl'lent
Phase I of the Autol'lated Array Assel'lbly Task
Si Production Process Evaluation
Coopositicn NeaSlJreYlents try Analyt...i.cal catalysis
Feasibility StudY of Solar Dolle Encapsulation of P\I Arrays
Mitre Solar Energy Test SystePi Evaluation
SLwort of the LifetiYle Cost Performnce (LCP) Modeling Effort
StudY to SUpport OeveloDl'lent of Solar Array Nufacturing Industry Costing Standards
SOlar Array Prolilction SysteA (SAPS) Production Progra'l Hi3"ldbook & Mathenatics Model
Phase 2 of the Array AYtol'lated Asset'lbly Task
AutOAation EQuij:flent Develoj:flent and Modification
Phase 2 of the Array AYtol'lated Assel'lbly Task
Use of Glass Reinforced Concrete (GRC) as a Substrate for PV Molilles
Vi tregraph Coaings on Mullite
PV Mol1Jle and Array Materials
SOlar ArraylNodule safety ReQuirefllents
SOlar Cells on ttllticrystalline Slbstrates of Si Refined try OS/PtIS Process
Sil~e to Si Process
OptiAization MethodS & Si Solar cell Nuyerical Models
Si Solar cell FabriC<ltioo Tecmology
Delayed fracture of Si
Devel~YK4nr of tlethods of Characterizing Si S~.eet Naterial
Si Sheet wi ttl Molecular BeaPi [pitaxy for Hi. -Efficieocy SOlar cells
Visible-Irradiaree Optical Pr~erties of O:Jnducting PolyRers
Tine Resolved Spectrosc~ic HeaSlJreYlents
SUrface and Allied Studies in Si SOlar cells
StudY of the Abrasive Wear Rate of Silicon
Stres~StrainAM~sisofSi~~
S~thetic Procelilres for PolYlleric
IN Stabilizers and Absorbers
Determne Effects of PresSlJre of Reactant Gases Where Si is in Contact wi ttl Die & Cootainer & Cootainer Metals
Analysis & EValuation of Processes ~d EQI.dj:flent
Hot formng of Si. Si Sheet Growth Oeveloj:flent
Analysis & Evaluation of I'blille Experil'lental Process S)lsteA OevelopYlent Unit Processes
Chairmn of Advisory Comittee on LOY Cost Solar cells
Si Solar cells of Near 20\ Efficiency
l4eb-Dendritic Ribbon Growth
OPOI
..
.....
..
..
IIREA
RES
RES
SINAT
OR
SHEET
SHEET
SINAT
PO
SINAT
SINAI
SHEET
PO
SHEET
PO
SINAI
SINAT
ESflP
PO
PMI
PMI
PO
PO
PO
..
..
..
.
..
..
..
...
..
PO
PO
SHEET
RES
RES
SINAT
SINAT
OR
SHEET
SHEET
SHEET
OR
OR
RES
OR
SHEET
SHEET
PO
SHEET
PO
SHEET
PO
SHEET
OR
SHEET
COMPANY
CD
(j)
University of Southern california
Uni versity of Southern california
University of Southern california
University of Toronto
University of Washirv;lton
Varian Associates
Varian Associates
Washington Universi tJ' at St. Louis
Washington Universi tJ' at St. Louis
westirv;lhouse Electric Corp.
westinghouse Electric Corp.
Westinghouse Electric Corp.
Westinghouse Electric Corp.
westinghouse Electric Corp.
westinghouse Electric Corp.
Westinghouse Electric Corp.
Westirv;lhouse Electric Corp.
westinghouse Electric Corp.
westinghouse Electric Corp.
westinghouse Electric Corp.
Westirv;lhouse Electric Corp.
Westinghouse Electric Corp.
Wilkes College
\.lyle Laboratories
Xerox Corp.
LEGEND:
Cz - CZochralski
DR - Device Rese-3rch
NPFA - I10dule Perfornalce r; Failure Analysis
OPEN - ~en or Continuing Contracts
PA&I - Project Analysis l; Integration
PO - Process Developrqent
PV - Photovoltaic(s)
RES - Reliability £; Engineering SCience
SHEET- Advancoo Silicon Sheet
Si - Silicon
SINAI- Silicon I1aterials
COIrTRACT TITLE
Neasurel'lent of the Absorption Coefficients of Si
Accelerated Aging and Analysis of PolYl'lers
Study of Deep Level Il'lPurities and Defects in Si
Photodegradation in Solar Cell Modules of Slbstrate & Superstrate Design
Si Inversion Layer Solar Gells & Carrier Li fetine Measurel'lents
Developl'lent of Advanced Methods for CZO Growth
Slicing of Si into Sheet t1aterial
Cz Crystal Modeling Study
FluidiZed Bed Reactors for Production of Si fran Silane
Dendritic Web-Type Solar Gell Nini-Modules
Developl'lent of Si Sheet by Dendritic IJeb Ribbon Process
Methods of Production of Large Areas of Si Sheet b'y 'the Web Dendrite Process
Si Dendritic LJeb Material Process DevelopFlent
Phase 2 of the Array Autonated Assel'lbly Task
Process for High Gapacity Arc Heater Prod of Si for Solar Arrays
Part 2-Develop & Define Requirenents of Purity for Solar Gell Grade Si
Process Research of non-Cz Si Naterial
Single Crystal 5i Dendritic Web Rbbon & Hi-Efficiency Solar Cells
Dendritic Web Process DevelopFlent
Laser Assisted Solar cell Metallization Processing
High-Efficiency Solar Cells on 5i Web
Process Researctl of Non-Cz Si Material
ElectroCheAical Research on Solar Cells aJ)j PV natetials
Durability/Reliability PerforPlance Criteri<l r; Test I1ethods for Array SUbsystel'l EleAents
Center Punched Solar cell Nodule Developrqent Effort
.,
OPal
..
..
.
..
.
..
..
...
.....
....
AREA
SHEET
PO
DR
RES
DR
SHEET
SHEET
SII1I\T
SII1I\T
RES
RES
SHEET
PO
PO
PO
PO
PO
SINI\T
SHEET
PO
DR
PO
RES
RES
PO
More Technology Advancements
Dendritic web silicon ribbons are grown to so/ar-{;ell
thickness. Progress is shown by experimental ribbons
grown in 1976 and 1978 and a ribbon grown in a
Westinghouse Electric Corporation pilot plant.
The edge-defined film-fed growth silicon ribbons are
grown to so/ar-{;ell thickness. A DOE/FSA-sponsored
research ribbon grown in 1976 is shown next to a
nine-sided ribbon grown in a Mobil Solar Energy
Corporation funded configuration.
FRAME
GLASS (STRUCTURAL}
SEAL
GASKET
Czochralski silicon crystals as grown are
sawed into thin circular wafers. (Support for
this effort was completed in 1981.)
Prototype modules have passed UL 790 Class A
burning brand tests which are more severe than
this spread of flame test.
Typical superstrate module design is shown with the
electrically interconnected solar cells embedded in a
laminate that is structurally supported by glass.
Materials and processes suitable for mass production
have been developed using this laminated design.
This 13.7 % efficiency prototype module was made
by Spire Corporation. If solar cells more recently
fabricated from float-zone silicon wafers were made
into a module, the module efficiency should be
about 15%.
Photovoltaic Applications
1975
u.s.
Coast Guard buoy
with photo voltaic-powered
navigational light.
Photo voltaic-powered corrosion protection
of underground pipes and wells.
Later...
House in Carlisle, Massachusetts, with a 7.3-kW
photo voltaic roof-top array. Excess photo voltaicgenerated power is sold to the utility. Power is
automatically supplied by the utility as needed.
A 28-kW array of solar cells for crop irrigation
during summer, and crop drying during winter
(a DOE/University of Nebraska cooperative project).
1985
1.2 MW of photo voltaic peaking-power generation
capacity for the Sacramento Municipal Utility District.
(The 8 x 16 ft panels are mounted on a north-south
axis for tracking the sun.)
JPL 400-279 10/85
(5101-279)